diff --git "a/attnserver.run_attnserver.slurm.sh.343215.err.log" "b/attnserver.run_attnserver.slurm.sh.343215.err.log" --- "a/attnserver.run_attnserver.slurm.sh.343215.err.log" +++ "b/attnserver.run_attnserver.slurm.sh.343215.err.log" @@ -12344,3 +12344,24885 @@ frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/ frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) [rank16]:[W621 21:58:14.945168437 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:15.270844425 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:15.274519297 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:15.270937759 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:15.274556614 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:15.158510139 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:15.158562050 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:15.330752121 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank9]:[W621 21:58:15.161857715 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:15.158534324 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:15.334465036 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:15.161874611 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:15.161895784 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:15.167603703 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:15.171652435 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:15.168023962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:15.171949462 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:15.265922673 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:15.344558396 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:15.265929376 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:15.348513289 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:15.346647267 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:58:15.269776350 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:15.265929086 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank10]:[W621 21:58:15.187159512 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:15.269793221 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:15.269810053 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:15.350560785 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:15.266167763 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:15.187168408 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:15.269923232 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:15.191917152 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:15.191921748 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:15.271433069 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:15.275196600 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:15.271827124 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:15.275752478 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:15.652135700 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:15.655812496 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:15.656112260 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:15.660040594 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:15.721626824 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:15.725556193 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:15.698585276 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:15.702064433 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:15.925523325 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:15.929582355 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:15.945326996 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:15.949022250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:16.274686934 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:16.278614066 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:16.274742750 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:16.278648922 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:16.162024650 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:16.334638779 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank9]:[W621 21:58:16.162002014 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:16.165451037 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:16.338329232 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:16.165475713 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:16.162047567 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:16.165501888 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:16.171853600 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:16.348699656 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:16.270025864 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:16.176062839 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:16.269951245 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:16.172254304 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:16.352529651 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:16.350735562 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:16.270044193 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:16.176198953 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank10]:[W621 21:58:16.192067712 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:16.354802272 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:16.273831850 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:16.273838265 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:16.269948707 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:16.192076408 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:16.273859641 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:16.273880566 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:16.275394519 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:16.196722340 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:16.196728195 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:16.279097155 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:16.275936694 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:16.279789932 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:16.656029513 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:16.659826613 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:16.660197908 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:16.664129247 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:16.725735026 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:16.729558228 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:16.702253268 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:16.706202254 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:16.929733359 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:16.933620646 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:16.949189241 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:16.952994458 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:17.278781253 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:17.282616784 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:17.278832455 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:17.282651506 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:17.165636385 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:17.165638496 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:17.338588266 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:17.342376921 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:17.169195566 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:17.165596048 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:17.169209594 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:17.169230134 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:17.176240834 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:17.180342280 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:17.176392019 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:58:17.274032644 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank30]:[W621 21:58:17.352695563 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:17.180378903 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:17.273983205 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:17.356480613 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:17.354961371 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:58:17.277922783 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:17.273979416 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:17.277952153 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:17.273979756 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +[rank27]:[W621 21:58:17.358954467 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:17.277969656 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:17.277995673 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:17.279294981 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:17.283108315 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:17.196863218 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank19]:[W621 21:58:17.279935063 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:17.196880470 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank19]:[W621 21:58:17.283834128 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:17.201525686 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:17.201534656 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:17.660036277 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:17.663707199 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:17.664282344 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:17.668046656 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:17.729724273 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:17.733608699 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:17.706362673 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:17.710116519 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:17.933770306 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:17.937742375 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:17.953156973 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:17.956734288 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:18.282766987 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:18.286621010 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:18.282813507 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:18.286657961 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:18.169370540 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:18.169358564 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank25]:[W621 21:58:18.342487155 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:18.173136223 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank9]:[W621 21:58:18.169352934 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:18.346083753 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:18.173171802 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:18.173190404 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:18.180599132 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:18.184731644 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:18.180628999 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:18.184774711 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:18.356648007 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:18.278090371 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:18.278097133 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:18.360410310 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:18.359117565 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:18.281883083 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:18.278130479 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:18.363033972 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:18.281898901 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:18.278090822 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:18.281914977 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:18.281937749 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:18.283312081 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:18.201666745 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:18.287036221 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:18.201674340 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank19]:[W621 21:58:18.284020485 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:18.206284967 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:18.206295231 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:18.287832548 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:18.663904363 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:18.667583250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:18.668182414 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:18.671974010 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:18.733778984 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:18.737661152 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:18.710290548 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:18.714200529 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:18.937909931 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:18.941943564 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:18.956897792 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:18.960568286 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:19.286820755 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:19.290713516 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:19.286788196 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:19.290749072 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:19.173332663 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:19.177080716 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:19.346195255 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:58:19.173333641 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:19.349906589 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:19.173333099 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:19.177122355 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:19.177140631 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:19.184936084 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:19.184935686 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:19.188954503 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:19.188985234 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:19.360560002 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:19.282012070 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:19.285888278 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:19.364473362 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:19.282090053 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:19.363195236 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:19.282056594 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:19.366949008 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:19.285926818 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:19.282086229 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:19.285944940 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:19.285969323 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:19.287246440 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:19.291001877 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:19.288023129 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:19.206424056 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:19.291858603 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:19.206444643 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:19.211203519 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:19.211219202 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:19.667782854 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:19.671617250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:19.672121752 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:19.676079145 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:19.737876889 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:19.741672609 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:19.714438581 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:19.718541852 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:19.942081142 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:19.946151672 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:19.960734286 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:19.964538515 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:20.290881283 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:20.294683197 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:20.290904597 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:20.294721502 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:20.177285243 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:20.181133850 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:20.177298717 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank25]:[W621 21:58:20.350009325 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:20.177300900 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:20.353617141 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:20.181181386 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:20.181198193 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:20.189124505 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:20.193062947 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:20.189132888 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:20.364610187 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:20.286007549 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:20.193100822 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:20.211341920 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:20.368411072 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:20.289725520 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:20.286079879 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:20.367120814 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:20.211374776 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:20.286097194 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:20.370758598 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:20.289763543 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:20.286107008 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank10]:[W621 21:58:20.216114758 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:20.216142624 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:20.289786510 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:20.289812097 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:20.291205202 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:20.294970519 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:20.292050622 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:20.295862241 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:20.671811688 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:20.675408367 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:20.676219050 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:20.679871453 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:20.741837990 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:20.745732281 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:20.718722468 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:20.722657645 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:20.946375325 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:20.950434256 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:20.964712250 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:20.968633471 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:21.294843332 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:21.298702025 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:21.294887288 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:21.298737265 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:21.181278642 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:21.185129151 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:21.181348788 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank25]:[W621 21:58:21.353744411 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:21.181348472 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:21.357465541 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:21.185176662 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:21.185193148 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:21.193228476 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:21.368543529 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:21.289840190 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:21.197143760 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:21.193261311 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank30]:[W621 21:58:21.372336849 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:21.293582061 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:21.289913474 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:21.370910931 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank11]:[W621 21:58:21.197195035 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:21.289943606 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:21.374478834 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:21.293619576 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:21.289933073 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:21.293641407 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:21.293665963 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:21.295187051 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:21.298841156 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:21.296037951 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:21.299714386 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:21.216253992 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:21.216293199 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:21.221041485 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:21.221062228 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:21.675616125 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:21.679434715 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:21.680121166 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:21.684055406 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:21.745908898 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:21.749873675 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:21.722826785 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:21.726650599 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:21.950553627 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:21.954380977 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:21.968801616 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:21.972702434 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:22.298866071 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:22.302648978 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:22.298884289 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:22.302685880 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:22.185261758 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:22.189163597 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:22.185329118 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank25]:[W621 21:58:22.357591898 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:22.185330902 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank25]:[W621 21:58:22.361331250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:22.189193262 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:22.189220723 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:22.197327551 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:22.293694287 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:22.201083067 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:22.372533228 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:22.197341126 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:22.297611204 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:22.293758968 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:22.376391097 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:22.201125175 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:22.374637219 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:22.293813196 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:22.378309180 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:22.297650129 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:22.293780675 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:22.297677267 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:22.297703144 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:22.299022196 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:22.302655580 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:22.299873614 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:22.221176997 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank19]:[W621 21:58:22.303536981 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:22.221205158 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:22.225741905 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:22.225756521 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:22.679639084 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:22.683249500 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:22.684222699 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:22.688146764 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:22.750043067 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:22.754048355 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:22.726840400 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:22.730836545 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:22.954514611 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:22.958421348 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:22.972877728 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:22.976723747 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:23.302824189 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:23.306701601 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:23.302821799 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:23.306799672 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:23.189316404 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:23.193298333 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:23.189364558 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank25]:[W621 21:58:23.361435752 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:23.189352621 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:23.365071254 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:23.193338691 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:23.193348067 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:23.201258074 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:23.205240596 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:23.201260028 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank30]:[W621 21:58:23.376526431 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:23.297719262 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:23.205285166 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:23.380148818 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:23.301357360 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:23.378482655 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:23.297809061 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:23.382114469 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:23.301496126 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:23.297824821 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:23.297837820 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:23.301591675 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:23.225883211 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:58:23.301614992 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank22]:[W621 21:58:23.302865690 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank13]:[W621 21:58:23.225907182 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:23.306540431 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank19]:[W621 21:58:23.303671596 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank10]:[W621 21:58:23.230685417 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:23.230702794 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:23.307339999 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:23.683452694 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:23.687185187 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:23.688313090 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:23.692386156 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:23.754220524 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:23.758075589 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:23.730984910 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:23.734588094 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:23.958555381 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:23.962521668 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:23.976878311 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:23.980611349 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:24.306857322 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:24.310685497 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:24.306959269 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:24.310722108 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:24.193444108 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:24.197257559 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:24.193489493 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:24.365232382 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:24.193488163 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:24.369051881 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:24.197291692 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:24.197315626 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:24.205377611 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:24.209175245 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:24.205450854 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:24.301471664 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:24.380351053 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:24.209271765 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:24.305251884 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:24.301611156 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:24.384139424 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:24.382267579 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:24.305289609 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:24.301730943 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:24.385867445 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:24.305319250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:24.301783179 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:24.305577127 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:24.306831756 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:24.310700747 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:24.307478396 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:24.311161134 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:24.230819305 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:24.230858436 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:24.235612115 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:24.235631096 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:24.687376725 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:24.691042054 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:24.692529062 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:24.696273620 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:24.758243234 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:24.762247543 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:24.734751346 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:24.738342157 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:24.962670832 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:24.966666141 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:24.980786814 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:24.984488374 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:25.310859668 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:25.314518093 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:25.310846993 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:25.314552972 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:25.197458696 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:25.369182187 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:58:25.200981741 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:25.197435576 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:25.197398938 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:25.372934433 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:25.305363855 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:25.384306943 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:25.309087239 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:25.201015777 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:25.201047926 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:25.305397171 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:58:25.209422468 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:25.388135928 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:25.386012786 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:25.305448460 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:25.213074282 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:25.209365069 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:25.389610584 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:25.309125885 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:25.309153141 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:25.213105292 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:25.305740046 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:25.309521669 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:25.310924075 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:25.314848211 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:25.311333518 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:25.235746062 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:25.235778440 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank19]:[W621 21:58:25.315306529 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:25.240334853 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:25.240341030 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:25.691229962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:25.694820945 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:25.696422985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:25.700219235 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:25.762417577 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:25.766385217 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:25.738502394 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:25.742156610 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:25.966820674 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:25.970730666 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:25.984651366 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:25.988328577 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:26.314678337 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:26.314706025 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:26.318401264 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:26.318429766 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:26.201196720 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:26.204743672 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:26.201174615 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:26.373096704 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:26.201120836 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:26.376975286 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:26.204778745 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:26.204803882 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:26.213271635 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:26.213271551 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:26.388315178 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:26.309199495 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:26.312956468 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:26.309240381 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank14]:[W621 21:58:26.217107522 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:26.217122952 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:26.392211392 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:26.389763534 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:26.309277183 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:26.393402632 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:26.312995654 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:26.313022794 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:26.309684036 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:26.313375704 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:26.315060575 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:26.318795350 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:26.315501049 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:26.319330842 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:26.240474593 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:26.240486118 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:26.244893215 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:26.244893212 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:26.695120167 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:26.698963277 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:26.700353402 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:26.704175094 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:26.766560296 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:26.770419521 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:26.742315275 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:26.745883660 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:26.970924763 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:26.975101921 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:26.988498670 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:26.992312642 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:27.318622822 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:27.322463795 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:27.318583038 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:27.322501318 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:27.204922174 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:27.208495508 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:27.204934240 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:27.377263462 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:27.204961219 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:27.381171132 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:27.208552907 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:27.208569064 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:27.217258389 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:27.313079760 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:27.221303548 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:27.217287260 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank30]:[W621 21:58:27.392367438 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:27.221337097 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:27.316677091 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:27.313108649 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank30]:[W621 21:58:27.396108295 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:27.393571870 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:27.316704859 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:58:27.313234041 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:27.397267756 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:27.316922607 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:27.313538722 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:27.317224009 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:27.319008210 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:27.322839292 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:27.319505053 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:27.323279878 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:27.245029441 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:27.245045920 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:27.249437082 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:27.249452552 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:27.699150566 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:27.702662833 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:27.704317284 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:27.708070429 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:27.770549314 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:27.774467975 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:27.746017872 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:27.749291799 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:27.975247118 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:27.979028809 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:27.992449313 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:27.996216696 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:28.322623096 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:28.326329575 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:28.322665724 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:28.326364195 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:28.208623863 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:28.212065317 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:28.381336010 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank9]:[W621 21:58:28.208663053 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:28.385184248 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:28.212163187 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:28.208693865 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:28.212191939 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:28.221470409 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:28.225436610 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:28.221501972 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:28.225465583 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:28.396257612 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:28.316803068 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:28.316811241 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:28.399879685 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:28.397425658 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:28.320328010 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:28.320347431 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:28.317153460 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:28.401055580 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:28.320806154 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:28.317356054 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:28.320871056 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:28.323078902 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:28.326951202 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:28.323483253 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:28.327294691 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:28.249571285 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:28.249604224 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:28.253999710 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:28.254020775 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:28.702848781 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:28.706397749 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:28.708218684 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:28.712133988 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:28.774609590 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:28.778581795 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:28.749454626 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:28.753220173 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:28.979198358 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:28.983175488 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:28.996360519 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:28.000056303 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:29.326500271 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:29.330306930 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:29.326530609 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:29.330344175 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:29.212181815 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:29.215965358 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:29.212288306 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:29.385364442 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:29.212323363 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:29.389183733 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:29.216017669 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:29.216044168 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:29.225657221 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:29.229444710 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:29.225613757 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:29.229480993 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:29.320443655 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:29.400044692 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:29.320459028 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:29.403632143 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:29.401230287 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:29.323987307 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:29.324019412 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:58:29.320985764 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:29.405024915 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:29.324506804 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:29.321030934 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:29.324678744 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:29.327167419 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:29.331008366 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:29.327515189 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:29.331365250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:29.254137642 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:29.258711347 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:29.254245119 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:29.258850573 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:29.706592922 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:29.710365140 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:29.712303340 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:29.716150972 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:29.778720603 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:29.782461708 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:29.753374255 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:29.757153706 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:29.983336098 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:29.987368983 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:29.000210200 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:29.004034587 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:30.330499052 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:30.334182920 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:30.330489135 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:30.334237408 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:30.216081031 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:30.219800345 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:30.216145148 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:30.219839774 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:30.216177624 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:30.389397310 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:30.393240141 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:30.219875813 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:30.229642672 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:30.324135983 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:30.233695261 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:30.229637432 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:30.403782841 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:30.324109751 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:30.407420016 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:30.233730831 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:30.405188927 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:30.327785570 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:30.327809199 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:30.324623297 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:30.409064230 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:30.328226315 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:30.324845283 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:30.328487602 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:30.331312446 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:30.335049158 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:30.331536928 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:30.335461561 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:30.258838189 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:30.263337209 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:30.259001233 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:30.263376830 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:30.710545360 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:30.714102436 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:30.716307817 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:30.720072368 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:30.782600316 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:30.786340429 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:30.757288519 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:30.760878148 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:30.987521874 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:30.991632630 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:30.004185314 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:30.007827559 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:31.334357000 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:31.338137084 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:31.334407044 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:31.338173264 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:31.219937543 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:31.223809865 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:31.219991308 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:31.220029348 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:31.393496546 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:31.223848531 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:31.223877622 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:31.397360698 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:31.233829016 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:31.237610495 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:31.233909889 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank18]:[W621 21:58:31.327901226 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:31.237655497 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:31.407603881 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:31.331447930 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:31.327941313 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank30]:[W621 21:58:31.411313945 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:31.331564481 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:31.409232061 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:58:31.328353836 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:31.413243630 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:31.331864536 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:31.328657523 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:31.332305365 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:31.335285524 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:31.339122614 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:31.335633661 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:31.339434014 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:31.263469721 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:31.263528040 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:31.268127727 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:31.268152466 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:31.714265018 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:31.717826907 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:31.720215298 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:31.724024958 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:31.786479760 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:31.790412713 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:31.761039983 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:31.765082718 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:31.991801783 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:31.995770250 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:31.008037128 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:31.011821993 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:32.338303381 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:32.342119287 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:32.338365930 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:32.342156659 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:32.224066893 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:32.223976394 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:32.227805065 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:32.397553728 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:32.227830593 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:32.224022686 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:32.401422936 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:32.227863927 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:32.237799720 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:32.237758910 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:32.331687564 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:58:32.241481700 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:32.241511772 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:32.411500330 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:32.331609664 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:32.415332700 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:32.335450066 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:32.335474064 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:32.413410986 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:58:32.332042113 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:32.417293545 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:32.335767231 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:32.332563591 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:32.336384355 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:32.339322561 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:32.268264309 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank22]:[W621 21:58:32.343063309 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank19]:[W621 21:58:32.339598513 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:32.268298273 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:32.343488720 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:32.272898544 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:32.272905362 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:32.717953023 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:32.721499430 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:32.724173739 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:32.727951475 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:32.790594913 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:32.794462801 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:32.765267425 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:32.769082899 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:32.995925690 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:32.999882265 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:32.011977014 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:32.015777132 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:33.342278659 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:33.346048361 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:33.342330137 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:33.346087198 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:33.228009978 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:33.228016296 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:33.231512903 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:33.227972074 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:33.401611259 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:33.231539930 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:33.231563267 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:33.405503960 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:33.241643536 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:33.245517766 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:33.241643554 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:33.415498862 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:33.335614066 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:33.339474686 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:33.335636673 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:33.245563122 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:33.419281103 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:33.417439450 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:33.339512682 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:33.335934681 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:33.421206618 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:33.339733979 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:33.336614135 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:33.273039997 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:33.273053811 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:58:33.340409869 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank22]:[W621 21:58:33.343249615 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank10]:[W621 21:58:33.277656993 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:33.277662203 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:33.346900480 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:33.343666055 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:33.347624140 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:33.721695173 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:33.725319311 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:33.728149750 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:33.731934688 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:33.794706103 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:33.798472566 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:33.769230081 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:33.772724424 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:33.000066433 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:33.004030412 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:33.015954414 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:33.019727296 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:34.346277214 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:34.350219146 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:34.346231974 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:34.350309144 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:34.231680627 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:34.235441560 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:34.231729581 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:34.231679772 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:34.405697556 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:34.409527415 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:34.235484043 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:34.235506935 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:34.245662061 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:34.249625758 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:34.245713488 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:34.419429184 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:34.339692817 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:34.339831750 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:34.249672986 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:34.423058627 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:34.421384965 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:34.343737109 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:34.339874020 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:34.425241216 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:34.343778661 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:34.343781710 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:34.340579007 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank10]:[W621 21:58:34.277799964 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:58:34.344359180 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:34.277810588 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank22]:[W621 21:58:34.347087667 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:34.351015617 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:34.282748543 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:34.282765939 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:34.347813090 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:34.351601866 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:34.725526962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:34.729478655 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:34.732094172 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:34.735935816 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:34.798624609 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:34.802330920 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:34.772887499 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:34.776554194 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:34.004196988 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:34.008203713 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:34.019899998 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:34.023509597 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:35.350389525 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:35.354124198 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:35.350464133 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:35.354162866 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:35.235627358 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:35.235652192 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:35.239211928 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:35.235577688 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank25]:[W621 21:58:35.409671580 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:35.239238164 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:35.239259714 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:35.413485463 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:35.249773153 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:35.253638095 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:35.249860903 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:35.423211836 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:35.343901277 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:35.253681592 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:35.426895147 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:35.347784130 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:35.343916374 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:35.282895581 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:35.425461215 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:35.282919900 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:35.343893899 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:35.429507106 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:35.347825161 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:35.287771048 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:35.287785650 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:35.347849427 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:35.344529658 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:35.348333821 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:35.351199086 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:35.355217532 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:35.351764866 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:35.355538225 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:35.729678032 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:35.733348716 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:35.736097904 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:35.740020411 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:35.802461946 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:35.806135781 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:35.776759032 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:35.780567889 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:35.008411500 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:35.012592233 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:35.023707288 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:35.027628683 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:36.354256527 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:36.357929252 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:36.354320837 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank12]:[W621 21:58:36.239412970 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank23]:[W621 21:58:36.347906650 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:36.239354795 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:36.357980577 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank25]:[W621 21:58:36.413619005 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:36.351753902 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:36.347964019 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:58:36.243037085 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:36.239490401 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:36.417221222 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank30]:[W621 21:58:36.427029364 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:36.243053287 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:36.348033115 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:36.243083479 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:36.253786313 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:36.430667201 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:36.351792237 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:36.429667163 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:36.351809240 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:36.257664237 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:36.253831189 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:58:36.348502389 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank27]:[W621 21:58:36.433373500 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:36.257707175 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:36.352289344 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:36.355426931 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:36.359431604 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:36.355699122 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:36.359843092 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:36.287907825 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:36.287942907 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:36.292708531 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:36.292726177 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:36.733620231 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:36.737195419 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:36.740176550 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:36.744149985 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:36.806248681 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:36.809918740 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:36.780731195 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:36.784184185 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:36.012754241 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:36.016858447 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:36.027847495 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:36.031858179 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:37.358098423 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:37.361825433 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:37.358151847 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:37.361862105 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:37.243187946 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:37.243241461 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:37.246908617 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:37.243214082 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:37.417354881 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:37.351868711 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:37.421008123 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:37.246932617 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:37.246950035 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:37.430780615 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:37.257807032 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:37.355618285 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:37.351940495 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:37.261623645 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:37.257898107 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:37.434570747 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:37.433546623 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:37.351955096 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:37.261688734 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:37.355659396 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:37.292874424 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:37.437540747 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:37.355683506 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:58:37.352482896 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:37.292897572 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:37.356224345 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:37.297693064 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:37.297705230 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:37.359609105 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:37.363518091 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:37.360063446 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:37.363777381 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:37.737370529 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:37.740861034 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:37.744300112 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:37.748327680 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:37.810024539 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:37.813821421 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:37.784341661 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:37.788123812 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:37.017021239 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:37.021088425 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:37.032054293 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:37.035889468 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:38.362021680 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:38.365890698 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:38.361997677 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:38.365927514 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:38.247031745 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:38.250749759 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:38.247120462 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:38.250780493 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:38.421143558 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:38.247107395 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:38.424752458 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:38.250820080 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:38.261762688 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:38.265784524 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:38.261838117 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:38.434675956 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:38.355736908 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:38.359522852 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:38.355811972 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:38.265829811 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:38.438283315 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:38.297839766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:38.437691594 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:38.355889941 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:38.297848982 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:38.441545380 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:38.359564767 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:38.359595810 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:38.356464228 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:58:38.302740853 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:38.302759815 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:38.360206789 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:38.363641248 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:38.367406067 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:38.363992549 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:38.367589424 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:38.741059250 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:38.744645633 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:38.748488445 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:38.752367067 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:38.813951730 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:38.817563665 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:38.788271521 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:38.791711811 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:38.021241779 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:38.025189056 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:38.036093083 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:38.039918638 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:39.366043466 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:39.369857858 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:39.366083073 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:39.369897738 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:39.250868352 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:39.250901330 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:39.254371681 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:39.254393507 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:39.424935062 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:39.250980641 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:39.428696677 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:39.254569681 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:39.265997941 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:39.438391274 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:39.359641025 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:39.270088786 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:39.266009190 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:39.363393562 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:39.359716623 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:39.441994884 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank27]:[W621 21:58:39.441697682 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:39.270127182 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:39.359745744 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:39.445690858 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:39.363438258 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:39.363458229 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:39.360376113 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:39.364039632 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:39.367554120 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:39.371372873 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:39.367914202 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:39.371623768 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:39.302881168 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:39.302914919 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:39.307838842 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:39.307856476 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:39.744859852 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:39.748659481 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:39.752529125 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:39.756434755 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:39.817711809 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:39.821418515 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:39.791859990 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:39.795376677 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:39.025325367 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:39.029231837 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:39.040087543 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:39.043647631 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:40.370010046 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:40.373820486 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:40.370049956 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:40.373856245 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:40.254498414 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:40.254549874 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:40.428832070 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank9]:[W621 21:58:40.257733721 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:40.257753940 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank15]:[W621 21:58:40.254683518 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:40.432481779 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:40.258122903 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:40.270285826 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:40.442095677 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:40.363512456 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:40.445675195 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:40.367502978 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:40.363606445 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank11]:[W621 21:58:40.274114811 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:40.270285969 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:40.445859573 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:40.274149667 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:40.363609000 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:40.449789662 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank10]:[W621 21:58:40.307978304 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:40.367552331 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:40.367572778 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:58:40.364194162 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:40.308017611 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:40.367806984 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:40.312769438 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:40.312772781 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:40.371505076 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:40.375439109 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:40.371805044 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:40.375575416 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:40.748850944 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:40.752569178 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:40.756580468 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:40.760459540 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:40.821584887 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:40.825403693 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:40.795553523 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:40.799411113 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:40.029352371 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:40.033220094 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:40.043815101 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:40.047440170 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:41.373980419 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:41.377869901 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:41.374036526 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:41.378005264 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:41.257898449 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:41.261402238 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:41.257863295 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:41.432746313 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:41.261433053 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:41.436530492 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:41.258233509 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:41.261549042 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:41.274273779 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:41.445804086 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:41.367613081 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:41.278236497 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:41.274414961 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:41.449438737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank27]:[W621 21:58:41.449944510 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:41.371517264 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:41.367689066 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:41.278271809 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:41.453559638 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:41.371568477 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:41.367743629 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:41.371596149 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:41.367961899 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:41.371626493 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:41.375559320 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:41.379492228 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:41.375759439 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:41.379692261 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:41.312912501 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:41.312911853 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:41.317572203 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:41.317580304 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:41.752760622 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:41.756254256 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:41.760620775 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:41.764444849 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:41.825625345 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:41.829542617 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:41.799557877 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:41.803376282 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:41.033344171 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:41.037101379 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:41.047666324 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:41.051146478 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:42.378041008 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:42.381975946 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:42.378191352 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:42.382038354 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:42.261570957 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:42.436701298 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:42.261685628 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:42.371644428 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:42.265175795 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:42.261568407 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank25]:[W621 21:58:42.440554191 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:42.449589632 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:42.375415160 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:42.265205180 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:42.265222234 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:42.453255333 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:42.371715566 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:58:42.278398237 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:42.453715440 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:42.282472573 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:42.278481042 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:42.375455934 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:42.371745587 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:42.457427489 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:42.282511343 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:42.371797182 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:42.375483571 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:42.375509483 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:42.379696196 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:42.383871780 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:42.379907425 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:42.383915371 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:42.317710209 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:42.317722225 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:42.322493285 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:42.322513615 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:42.756445573 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:42.759958337 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:42.764600716 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:42.768303947 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:42.829715726 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:42.833581208 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:42.803545271 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:42.807366820 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:42.037220066 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:42.040939317 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:42.051298999 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:42.054857402 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:43.382132273 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:43.385895203 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:43.382192973 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:43.385934863 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:43.265341672 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:43.265310338 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:43.268621297 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:43.268648997 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:43.265368680 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:43.440721017 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:43.268794157 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:43.444427888 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:43.282639451 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:43.453403633 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:43.375523065 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:58:43.286492973 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:43.282706167 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:43.379220453 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:43.286526142 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:43.457232557 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:43.375590053 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:43.457579374 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:43.379279146 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:43.461267483 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:43.375685861 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:43.379476880 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:43.375726376 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:43.379509620 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:43.384090599 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:43.388154289 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:43.384137822 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:43.388190802 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:43.322631704 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:43.322660170 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:43.327270109 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:43.327276410 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:43.760149216 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:43.763826923 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:43.768447960 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:43.772261878 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:43.833751277 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:43.837717386 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:43.807530115 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:43.811423511 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:43.041093938 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:43.045006387 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:43.055030981 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:43.058643389 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:44.386051863 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:44.389924288 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:44.386097526 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:44.389962621 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:44.268770978 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:44.272808410 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:44.268826612 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:44.444603405 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:44.268932052 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:44.272851825 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:44.448520739 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:44.272868744 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:44.286647803 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:44.457404282 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:44.379345841 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:44.379411260 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:58:44.290450161 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:44.461352909 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:44.286714856 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:44.461426735 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:44.383224280 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:44.383251149 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:44.379648152 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:44.290730505 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:44.465136694 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:44.383412457 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:44.379717279 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:44.383497872 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:44.388473190 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:44.388472748 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:44.392286316 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:44.392317425 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:44.327421084 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:44.327453744 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:44.332012453 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:44.332038357 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:44.764041540 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:44.767567962 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:44.772401742 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:44.776376873 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:44.837885798 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:44.841781681 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:44.811617697 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:44.815493341 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:44.045144375 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:44.049051270 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:44.058823567 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:44.062498576 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:45.390072604 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:45.393863489 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:45.390122797 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:45.393898510 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:45.272992395 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:45.276704010 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:45.448654968 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:45.383384790 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:58:45.272995722 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:45.383371338 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:45.272968026 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:45.452323918 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:45.461519820 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:45.387312434 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:45.383613798 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:58:45.276743712 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:45.276759859 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank11]:[W621 21:58:45.290627230 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:45.465316383 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:45.387329385 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:45.387350626 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:45.465275219 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:58:45.383707980 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:45.294504445 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:45.290870135 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank27]:[W621 21:58:45.468958546 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:45.387511297 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:45.294536475 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:45.392508937 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank13]:[W621 21:58:45.332218331 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:45.392563567 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:45.332218381 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:45.396269342 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:45.396281016 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:45.336719457 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:45.336734358 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:45.767776543 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:45.771326467 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:45.776518612 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:45.780258461 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:45.842018096 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:45.845787662 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:45.815645625 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:45.819375570 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:45.049189708 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:45.052961968 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:45.062675683 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:45.066406067 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:46.394043073 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:46.397808828 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:46.394003493 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:46.397840291 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:46.276885998 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:46.276881893 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:46.280655366 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:46.276889252 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank25]:[W621 21:58:46.452490215 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:46.280678046 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:46.280699241 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:46.456263146 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:46.294690096 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:46.465475762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:46.387453499 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:46.294671132 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:46.387519290 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:46.469316841 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:46.469139303 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:46.298350054 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:46.298378223 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:58:46.391204429 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:46.387471810 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:46.473134619 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:46.391232204 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:46.391256907 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:46.387706625 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:46.391376527 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:46.396484619 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:46.396471866 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:46.400230348 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:46.400249527 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:46.336876710 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:46.336876621 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:46.341582279 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:46.341600525 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:46.771503431 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:46.774985986 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:46.780400170 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:46.784214693 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:46.845931136 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:46.849596007 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:46.819554925 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:46.823728917 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:46.053089789 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:46.056691552 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:46.066573613 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:46.070007609 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:47.397968071 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:47.401827896 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:47.397996581 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:47.401864343 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:47.280842347 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:47.280842755 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:47.284659467 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:47.280824167 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:58:47.456436181 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:47.284674165 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:47.284694724 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:47.460324978 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:47.298521569 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:47.298522113 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:47.469484482 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:47.391352038 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:58:47.302147763 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:47.302172781 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:47.473285910 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:47.395093490 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:47.391384637 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:47.473285987 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:47.391387890 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:47.477045909 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:47.395126110 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:47.391525817 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:47.395150081 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:47.395168727 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:47.400503204 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:47.400558430 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:47.404340330 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:47.404362976 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:47.341762684 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:47.341762389 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:47.346305529 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:47.346316956 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:47.775163349 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:47.778648519 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:47.784358917 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:47.788166232 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:47.849732365 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:47.853702516 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:47.823883863 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:47.827355904 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:47.056798702 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:47.060292486 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:47.070159926 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:47.073705738 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:48.402018229 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:48.405726993 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:48.401987178 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:48.405761201 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:48.284815460 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:48.284806781 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:48.288393888 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:48.284859042 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:48.288421205 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:48.288437968 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:48.460494923 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:48.464246930 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:48.302338343 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:48.473476667 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:48.395258530 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:48.306297490 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:48.302442409 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:48.398916393 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:48.477302570 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:48.395274177 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:48.477197008 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:48.306355948 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:48.395240306 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:48.480850363 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:48.398953278 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:48.395331482 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:48.398978243 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:48.399001319 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:48.404573591 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:48.408573476 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:48.404573735 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:48.408749788 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:48.346450016 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:48.346466728 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:48.350853059 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:48.350859063 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:48.778825873 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:48.782358968 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:48.788315421 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:48.792105491 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:48.853857813 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:48.857586687 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:48.827540274 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:48.831502750 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:48.060408830 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:48.064304875 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:48.073880137 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:48.077481598 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:49.405926122 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:49.409889777 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:49.405914246 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:49.409925424 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:49.288583944 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:49.288590028 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:49.292269377 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:49.288560686 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank25]:[W621 21:58:49.464420697 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:49.292278742 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank9]:[W621 21:58:49.292304231 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:49.468276799 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:49.306507381 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:49.306474262 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:49.477442699 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:49.399038889 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:49.310347648 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:49.310373600 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:49.402829382 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:49.399142299 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:49.481239867 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:49.481016420 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:49.399109940 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:49.484725921 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:49.402865574 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:49.399126050 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:49.402891641 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:49.402911394 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:49.408791404 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:49.412633281 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:49.408928142 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:49.412669548 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:49.350992100 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:49.351001648 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:49.355319209 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:49.355319410 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:49.782562142 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:49.786178776 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:49.792248916 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:49.795992498 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:49.857761972 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:49.861726126 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:49.831666124 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:49.835547338 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:49.064426207 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:49.068060299 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:49.077623073 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:49.081288090 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:50.410044681 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:50.410086896 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:50.413902272 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:50.413933426 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:50.292409766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:50.296037015 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:50.292425987 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:50.292450890 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank25]:[W621 21:58:50.468540968 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:50.296089340 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:50.296107300 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:50.472381231 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:50.310549003 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:50.310549145 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:50.481373530 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:50.402936468 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:50.314480517 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:50.314506027 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:50.485024641 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:50.406578971 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:50.355473245 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:50.484857116 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:50.403059007 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:50.355473218 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:50.488150246 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:50.406655812 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:50.403039721 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank13]:[W621 21:58:50.360294227 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:50.360294217 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:50.406692449 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:50.403069289 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:50.406805823 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:50.412817170 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:50.416517110 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:50.412833931 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:50.416551181 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:50.786442326 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:50.790025970 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:50.796145101 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:50.799861614 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:50.861861833 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:50.865470000 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:50.835735137 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:50.839881065 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:50.068199588 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:50.072053015 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:50.081475598 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:50.085236070 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:51.414069263 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:51.414098227 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:51.417904950 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:51.417932957 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:51.296160525 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:51.472558355 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:58:51.299815120 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:51.296231299 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:51.406715328 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:51.476437720 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:51.406814005 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:58:51.296282835 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:51.485198362 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:51.299859171 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:51.299887760 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:51.410453550 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:51.410476018 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:51.314632833 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:51.488978145 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:51.406830773 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:51.488272869 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:51.314661930 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:51.491549130 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:51.410623502 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:51.406973893 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:51.318518148 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:51.318539816 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:51.410717836 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:51.416663394 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:51.420324399 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:51.416756636 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:51.420367378 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:51.360480426 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:51.360480541 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:51.364833881 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:51.364837715 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:51.790197293 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:51.793489785 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:51.800054914 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:51.803872496 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:51.865581571 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:51.869179898 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:51.840077251 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:51.843798291 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:51.072165575 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:51.075600294 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:51.085392985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:51.089083419 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:52.418054178 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:52.418116174 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:52.421823627 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:52.421851224 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:52.300049846 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:52.300062576 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:52.303518667 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:52.303544644 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:52.300045701 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank25]:[W621 21:58:52.476613970 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank15]:[W621 21:58:52.303653007 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:52.480473555 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:52.318683948 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:52.318684212 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:52.489122067 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:52.410567523 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:52.414224442 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:52.410620553 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank11]:[W621 21:58:52.322575737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:52.322601501 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:52.492775991 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:52.491682293 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:52.414255701 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:52.410749624 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:52.494965470 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:52.414412730 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:52.410881359 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:52.414612610 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:52.420582255 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:52.424359554 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:52.420541241 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:52.424397246 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:52.364965873 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:52.364967311 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:52.369545349 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:52.369548439 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:52.793696743 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:52.797071600 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:52.804025495 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:52.807660407 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:52.869287287 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:52.872867227 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:52.844017395 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:52.848082566 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:52.075769246 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:52.079654485 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:52.089246539 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:52.092876455 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:53.421989725 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:53.422048008 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:53.425985549 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:53.426013252 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:53.303691543 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:53.303692956 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:53.307537687 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:53.303834252 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:53.307552876 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:53.307572816 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:53.480629430 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:53.484591013 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:53.322740864 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:53.414402739 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:53.492915220 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:53.418073025 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:53.326709417 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:53.322741147 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:53.414395809 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:53.496604749 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:53.326756341 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:53.495085233 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:53.418203117 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:58:53.414567577 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:53.498435756 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:53.418318153 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:53.414794502 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:53.418576168 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:53.424624062 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:53.428517394 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:53.424583758 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:53.428552447 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:53.369693533 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:53.369693414 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:53.374084967 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:53.374087399 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:53.797232795 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:53.800508297 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:53.807801599 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:53.811757124 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:53.872997069 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:53.876756887 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:53.848268536 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:53.852245147 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:53.079778805 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:53.083589418 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:53.093036733 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:54.096691647 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:54.426156892 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:54.426198782 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:54.430104405 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:54.430125655 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:54.307704665 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:54.307721088 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:54.311654103 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:54.307752283 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:54.311677549 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:54.311690680 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:54.484768708 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:54.488695746 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:54.326848490 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:54.496739380 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:54.418337811 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:54.330657862 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:54.326895580 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:54.418284260 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:54.330688075 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:54.500571960 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:54.498563320 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:54.422067081 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:54.422085331 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:54.501878751 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:54.418451377 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:54.422146761 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:54.418743527 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:54.422495549 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:54.428729125 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:54.428757413 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:54.432669777 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:54.432699864 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:54.374217800 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:54.374218188 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:54.378565953 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:54.378570588 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:54.800691560 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:54.804418330 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:54.811901348 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:54.815828811 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:54.876893024 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:54.880614076 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:54.852407793 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:54.855847696 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:54.083716483 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:54.087505495 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:55.096855962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:55.100520478 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:55.430310055 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:55.434354667 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:55.430309984 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:55.434394032 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:55.311798688 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:55.311824585 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:55.315636358 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:55.311843277 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:55.315660528 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:55.315675246 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:55.488833679 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:55.492717847 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:55.330817145 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:58:55.422183947 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:55.500705153 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:55.334598407 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:55.330826033 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:55.425905897 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:55.504613550 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:55.422278963 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:55.334633008 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:55.502016952 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:55.422309862 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:55.505370263 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:55.426075775 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:55.426093456 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:55.422662618 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:55.426453761 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:55.432917094 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:55.436684630 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:55.432917238 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:55.436720373 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:55.378693932 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:55.378694762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:55.383047358 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:55.383052085 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:55.804602988 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:55.808048662 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:55.815981742 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:55.819787975 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:55.880759830 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:55.884680100 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:55.856053279 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:55.860017732 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:55.087625969 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:55.091332309 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:56.100680180 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:56.104329580 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:56.434531187 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:56.438529845 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:56.434619005 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:56.438591138 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:56.315780537 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:56.319590403 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:56.315840984 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:58:56.315914839 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:56.319638991 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:56.319653714 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:56.492853033 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:56.496664945 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:56.334731395 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:56.338468621 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:56.334806602 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:58:56.504796300 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:56.426119854 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:56.429941108 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:56.426263131 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:58:56.338541530 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:56.508739528 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:58:56.505513128 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:56.426255629 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank27]:[W621 21:58:56.508921822 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:56.430052508 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:56.430071629 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:56.426668848 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:56.430316333 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:56.436941498 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:56.440900814 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:56.436900528 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:56.440941645 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:56.383185153 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:56.383188524 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:56.387570113 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:56.387576918 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:56.808216585 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:56.811414798 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:56.819926019 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:56.823693028 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:56.884828294 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:56.888750847 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:56.860194758 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:56.863771487 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:56.091471017 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:56.095372104 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:57.104503826 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:57.108318526 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:57.438700274 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:57.442554357 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:57.438758835 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:57.442594983 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:57.319756070 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:57.319789631 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:57.323741671 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:57.319938988 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:57.323768197 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:57.323782704 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:57.496795681 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:57.500672898 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:58:57.338612417 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:57.342485923 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:57.509051499 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:58:57.430105323 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:57.338799497 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:57.512472737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:57.433924815 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:57.508910489 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:58:57.430218384 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:57.342765009 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:57.430222673 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:58:57.512873586 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:57.433993022 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:57.434014307 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:57.430477876 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:57.434224301 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:57.441183763 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:57.444969388 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:57.441116412 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:57.445008455 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:57.387710064 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:57.387718405 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:57.392074296 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:57.392076606 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:57.811572342 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:57.814778825 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:57.823832638 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:57.827673606 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:57.888923669 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:57.892846619 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:57.863929842 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:57.867623020 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:57.095497094 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:58.099296460 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:58.108478261 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:58.112146070 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:58.442703945 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:58.446530316 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:58.442789886 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:58.446570351 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:58:58.323907479 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:58.323890000 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:58.327694134 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:58.323961875 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:58.327711764 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:58.327728577 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:58.500826333 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:58.504773239 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:58.512587114 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:58.342637930 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:58:58.434064142 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:58:58.515805049 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:58:58.513039762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:58.434150830 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:58.342981855 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:58.346762215 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:58.346788506 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:58:58.516915776 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:58.437753581 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:58.437780272 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:58.434160538 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:58.437843278 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:58.434451026 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:58.438156220 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:58.445260635 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:58.449198384 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:58.445252805 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:58.449234201 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:58.392260692 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:58.392260838 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:58.396688476 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:58.396710246 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:58.814953832 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:58.818259361 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:58.827815769 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:58.831707358 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:58.893018581 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:58.896852376 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:58.867803294 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:58.871568801 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:58:59.099431921 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:58:59.103366383 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:58:59.112312008 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:58:59.115989448 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:58:59.446663184 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:58:59.450598642 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:58:59.446746251 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:58:59.450635661 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:58:59.327860103 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:59.327853456 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:58:59.331599128 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:59.327904482 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:58:59.331620652 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:58:59.331634128 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:58:59.504949271 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:58:59.509043201 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:58:59.515918384 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:58:59.346903101 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:58:59.519197632 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:59.437868140 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:59.517083270 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:58:59.437892396 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:58:59.346940325 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:58:59.520893806 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:58:59.350749997 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:58:59.350776073 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:58:59.441518603 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:58:59.441543262 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:58:59.437985313 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:58:59.441628584 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:58:59.438322251 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:58:59.442066148 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:58:59.449455728 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:58:59.453237362 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:58:59.449419134 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:58:59.453281286 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:58:59.396840587 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:58:59.396840159 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:58:59.401428713 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:58:59.401433728 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:58:59.818451407 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:58:59.822098850 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:58:59.831851952 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:58:59.835856179 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:58:59.897048020 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:58:59.901005168 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:58:59.871748649 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:58:59.875481496 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:00.103492857 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:00.107322796 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:00.116164000 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:00.119810454 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:00.450726293 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:00.454605308 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:00.450803665 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:00.454645808 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:00.331758504 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:00.335641414 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:00.331796238 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:00.331795970 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:00.335674802 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:00.335694617 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:00.509220266 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:00.513153801 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:00.350926763 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:59:00.519343413 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:00.441661302 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:00.354668629 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:00.522853226 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:00.445318706 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:00.441687495 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank11]:[W621 21:59:00.350915930 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:00.521047196 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:00.441765121 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:00.354771495 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:00.524720405 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:00.445354726 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:00.445374301 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:00.442241316 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:00.445952219 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:00.453490259 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:00.457282883 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:00.453641356 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:00.457594553 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:00.401617274 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:00.401617115 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:00.406237026 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:00.406237050 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:00.822289766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:00.825746404 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:00.836016502 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:00.839884182 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:00.901136799 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:00.904780149 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:00.875660755 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:00.879356837 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:01.107471985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:01.111483606 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:01.120046741 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:01.123833849 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:01.454745452 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:01.458631118 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:01.454837213 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:01.458716723 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:01.335882597 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:01.335852167 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:01.339632925 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:01.335787890 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:01.339656677 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:01.339675584 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:01.513330868 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:01.517362450 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:01.523003315 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:01.354804706 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:01.445468291 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:59:01.526597049 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:01.524858467 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:01.445521057 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:01.358675191 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:01.354909443 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:01.528507777 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:01.449110750 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:01.449135877 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:01.445541192 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank11]:[W621 21:59:01.358709341 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:01.449302671 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:01.446132058 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:01.449845972 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:01.457495541 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:01.457803731 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:01.461310857 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:01.461341898 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:01.406373892 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:01.406386816 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:01.410977382 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:01.410980178 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:01.825916615 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:01.829331370 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:01.840034571 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:01.843648672 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:01.904935385 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:01.908761217 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:01.879555839 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:01.883645296 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:02.111620616 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:02.115534947 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:02.124010079 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:02.127672892 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:02.458807263 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:02.462855975 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:02.458918060 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:02.462904084 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:02.339815103 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:02.339814476 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:02.343633418 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:02.339817002 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:02.343650501 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:02.343666579 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:02.517533277 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:02.521429685 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:02.526738757 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:02.530290963 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:02.449312433 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:02.358860087 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:02.528645542 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:02.449458205 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:02.362800266 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:02.532306200 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:02.358860128 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:02.453036550 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:02.449322375 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:02.453053633 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:02.453070830 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:02.362984424 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:02.450045279 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:02.453713022 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:02.461533230 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:02.465560219 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:02.461594854 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:02.465597308 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:02.411187171 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:02.411187413 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:02.415831658 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:02.415831655 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:02.829521940 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:02.833085272 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:02.843789676 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:02.847582504 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:02.908918989 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:02.912836737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:02.883847284 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:02.887977206 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:03.115681152 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:03.119638355 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:03.127862600 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:03.131573448 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:03.463029799 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:03.463099471 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:03.467022056 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:03.467050217 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:03.343815538 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:03.343846725 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:03.347706707 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:03.343816544 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:03.347721216 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:03.347744196 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:03.521607106 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:03.525552999 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:03.530455651 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:03.533814035 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:03.453202814 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:03.363144769 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:03.532461565 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:03.456834202 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:03.453219893 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:03.362970385 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:03.536342865 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:03.456976867 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:03.366967249 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:03.366993439 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:03.453396209 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:03.457147194 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:03.453881460 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:03.457590651 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:03.465776848 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:03.469556144 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:03.465828580 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:03.469612722 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:03.416026966 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:03.420775377 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:03.416076729 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:03.420811690 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:03.833254480 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:03.836621210 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:03.847727788 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:03.851565011 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:03.912993015 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:03.916900035 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:03.888168136 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:03.891972738 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:04.119788198 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:04.123833332 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:04.131747274 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:04.135427610 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:04.467147178 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:04.470856687 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:04.467239171 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:04.470902130 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:04.347913435 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:04.347915298 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:04.351854916 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:04.347872537 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank25]:[W621 21:59:04.525728201 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:04.351879993 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:04.351903459 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:04.529566096 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:04.533970044 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:04.367143056 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:04.456997353 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:59:04.537541957 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:04.536479493 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:04.457151815 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:04.367136284 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:04.371106321 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:04.540112840 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:04.461034338 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:04.461057623 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:04.457374982 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank14]:[W621 21:59:04.371132812 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:04.461185570 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:04.457782203 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:04.461558017 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:04.469772287 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:04.469815330 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:04.473836095 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:04.473868287 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:04.420907530 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:04.420954909 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:04.425727690 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:04.425736736 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:04.836792599 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:04.840355729 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:04.851707281 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:04.855516299 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:04.917052934 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:04.921042394 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:04.892178328 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:04.896279669 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:05.123969171 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:05.127871823 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:05.135596296 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:05.139258607 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:05.471021873 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:05.475026955 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:05.471109007 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:05.475065443 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:05.352085402 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:05.352115806 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:05.355990532 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:05.352143882 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:05.356025600 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:05.356048044 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:05.529747520 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:05.533623833 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:05.537678297 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:05.541028118 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:05.461160855 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:05.371263703 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:05.540249373 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:05.461220804 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:05.371258739 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:05.543972892 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:05.464968453 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:05.464998169 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:05.375035459 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:05.375062810 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:05.461402259 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:05.465219069 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:05.461740492 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:05.465538752 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:05.474032863 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:05.477868407 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:05.474097361 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:05.477907996 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:05.425865613 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:05.425884105 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:05.430593364 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:05.430597449 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:05.840520188 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:05.843887587 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:05.855666120 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:05.859509670 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:05.921189504 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:05.925350871 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:05.896489339 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:05.900466983 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:06.128008043 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:06.131793099 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:06.139417343 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:06.143112112 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:06.475138428 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:06.478898737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:06.475252203 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:06.479034212 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:06.356156664 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:06.359802331 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:06.356211287 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:06.356329647 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:06.359866358 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:06.359882541 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:06.533833120 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:06.537685563 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:06.541166187 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:06.544527075 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:06.465095625 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:06.544083818 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:06.465159990 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:06.375168234 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:06.547706238 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:06.468813883 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:06.468840271 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:06.378936562 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:06.375197526 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:06.465363565 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:06.378975154 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:06.468975291 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:06.465720486 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:06.469456296 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:06.478048690 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:06.481926102 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:06.478120646 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:06.481971146 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:06.430727720 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:06.430738784 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:06.435467412 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:06.435472718 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:06.844056294 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:06.847581484 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:06.859724037 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:06.863613062 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:06.925522125 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:06.929357367 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:06.900664614 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:06.904690002 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:07.131928809 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:07.135836815 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:07.143310504 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:07.147000236 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:07.479028098 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:07.482960944 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:07.479269565 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:07.483207155 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:07.359936935 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:07.363583663 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:07.360057260 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:07.363638719 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:07.360076950 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:07.363759716 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:07.537863850 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:07.541758528 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:07.544666651 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:07.548063730 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:07.469002247 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:07.379059629 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:07.547818489 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:59:07.472806737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:07.469017386 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank14]:[W621 21:59:07.382791373 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:07.379135111 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:07.551443828 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:07.472840671 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +[rank20]:[W621 21:59:07.469144378 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:07.382828296 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:07.472902608 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:07.469643447 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:07.473390205 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:07.482185157 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:07.486091488 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:07.482129691 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:07.486133761 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:07.435598699 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:07.435614950 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:07.440328172 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:07.440334953 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:07.847750700 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:07.851159542 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:07.863753653 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:07.867619965 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:07.929525803 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:07.933290379 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:07.904901680 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:07.908846503 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:08.135985037 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:08.139977503 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:08.147175101 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:08.150969620 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:08.483074445 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:08.486923903 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:08.483364277 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:08.487118091 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:08.363723502 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:08.363823339 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:08.367500154 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:08.363923449 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:08.367525730 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:08.367542432 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:08.541895175 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:08.545557153 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:08.548191944 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:08.551773344 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:08.472991154 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:08.551584766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:08.382935618 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:08.473125497 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:08.386820935 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:08.382977503 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:59:08.555219883 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:08.476856093 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:08.476876630 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank18]:[W621 21:59:08.472990293 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:08.386855995 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:08.476940779 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:08.473565383 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:08.477258895 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:08.486287990 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:08.490197985 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:08.486322666 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:08.490236079 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:08.440466288 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:08.440482552 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:08.445162148 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:08.445168662 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:08.851314751 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:08.854898231 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:08.867774842 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:08.871515883 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:08.933445490 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:08.937098961 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:08.909055879 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:08.912999295 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:09.140109967 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:09.143965272 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:09.151135732 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:09.154933914 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:09.487036397 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:09.490901117 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:09.487279166 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:09.491031509 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:09.367644054 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:09.367719367 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:09.371269619 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:09.367676762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:59:09.545729677 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:09.371285968 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:09.371306584 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:09.549570806 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:09.551906980 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:09.477081311 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:09.386954644 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:09.555296305 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:09.480796069 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:09.476996685 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:09.390760280 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:09.386998046 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +[rank30]:[W621 21:59:09.555335961 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:09.558995470 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:09.477018614 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:09.390802251 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:09.480837838 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:09.480863155 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:09.477425141 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:09.481149207 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:09.490403560 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:09.494329483 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:09.490403320 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:09.494368066 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:09.445294628 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:09.445311399 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:09.449803141 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:09.449817261 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:09.855082380 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:09.858907333 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:09.871670956 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:09.875519804 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:09.937254555 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:09.940991094 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:09.913323910 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:09.917350362 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:10.144107207 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:10.147974951 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:10.155112363 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:10.158692962 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:10.491042718 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:10.494977118 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:10.491185354 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:10.495021096 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:10.371443672 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:10.371435365 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:10.375041343 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:10.371449816 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:10.375065590 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:10.375080258 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:10.549717719 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:10.553358333 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:10.390948590 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:59:10.555458328 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:10.480948747 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:10.394907169 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:10.390938968 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:10.484748351 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:10.480999583 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:59:10.558975218 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank30]:[W621 21:59:10.559147315 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:10.395089909 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:10.484782782 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:10.481081964 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:10.562946935 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:10.484925184 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:10.481313798 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:10.485098937 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:10.494574366 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:10.498455023 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:10.494542486 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:10.498497534 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:10.449945886 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:10.449965754 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:10.454614564 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:10.454622744 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:10.859106065 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:10.862839744 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:10.875658279 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:10.879442498 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:10.941138743 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:10.944762280 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:10.917543599 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:10.921517688 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:11.148115399 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:11.152116495 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:11.158875040 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:11.162484630 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:11.495090915 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:11.499001269 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:11.495163791 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:11.499056661 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:11.375188433 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:11.375205821 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:11.378793749 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:11.378815758 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:11.375261932 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:59:11.553524457 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:11.557310703 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:11.378868980 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:11.559125658 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:11.562617819 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:11.563109086 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:11.484949026 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:11.395066799 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:11.566910973 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:11.488696659 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:11.398941205 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:11.484914177 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:11.395262222 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:11.488728148 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:11.485279692 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:11.399125556 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:11.489028000 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:11.485067275 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:11.489099909 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:11.498623300 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:11.502244072 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:11.498698489 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:11.502285182 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:11.454738221 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:11.454762859 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:11.459132222 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:11.459140802 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:11.863038087 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:11.866599730 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:11.879600384 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:11.883248657 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:11.944914266 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:11.948760017 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:11.921675574 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:11.925444284 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:12.152286593 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:12.156323515 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:12.162671361 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:12.166371050 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:12.499100541 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:12.502792922 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:12.499183905 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:12.502843804 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:12.378930094 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:12.378998880 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:12.557489209 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:59:12.382470928 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:12.382501784 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:12.379051492 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:12.561303917 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:12.382577562 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:12.562765646 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:59:12.488822469 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:12.399103847 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:59:12.566217211 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:12.492509596 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:12.567078946 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:12.488893015 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:12.403044956 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:12.399364324 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:12.570884427 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:12.492559580 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:12.489234861 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:12.403372929 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:12.492738700 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:12.489204033 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:12.492889519 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:12.502439634 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:12.506238230 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:12.502633711 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:12.506486910 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:12.459271821 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:12.459279659 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:12.463917244 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:12.463932387 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:12.866799888 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:12.870407058 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:12.883393844 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:12.887064836 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:12.948928036 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:12.952644725 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:12.925611063 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:12.929542698 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:13.156492910 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:13.160444700 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:13.166549302 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:13.170251339 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:13.502890610 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:13.506504215 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:13.502964534 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:13.506546205 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:13.382621871 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:13.382637211 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:13.386233237 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:13.382755855 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:13.386256801 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:13.386275576 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:13.561467138 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:13.565153866 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:13.566354710 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:13.569734058 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:13.571044642 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:13.574756037 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:13.492645466 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:13.403208288 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:13.496314925 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:13.492873977 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:13.407300590 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:13.403562067 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:13.496528662 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:13.407346639 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:13.493061217 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:13.496719495 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:13.492725336 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:13.496762411 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:13.506420882 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:13.510132578 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:13.506609137 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:13.510261983 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:13.464084107 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:13.464083958 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:13.468710054 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:13.468727277 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:13.870644171 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:13.874237460 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:13.887204419 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:13.890838040 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:13.952810818 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:13.956477377 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:13.929709902 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:13.933720596 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:14.160741149 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:14.164615512 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:14.170421840 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:14.174113862 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:14.506615282 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:14.510355961 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:14.506671281 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:14.510408357 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:14.386374662 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:14.386396652 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:14.390089264 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:14.386457555 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:14.390114271 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:14.390136543 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:14.565364480 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:14.569142283 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:14.569879489 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:14.573346293 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:14.496430194 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:14.407465207 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:14.574919813 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:14.500031456 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:14.411484418 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:14.407528711 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:14.578620533 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:14.496707448 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:14.411525396 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:14.500362092 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:14.496973373 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:14.496887639 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:14.500656470 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:14.500664990 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:14.510317143 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:14.513865007 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:14.510380270 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:14.514015774 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:14.468852052 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:14.468871750 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:14.473510429 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:14.473514806 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:14.874432827 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:14.878003835 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:14.890978444 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:14.894652216 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:14.956656605 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:14.960408006 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:14.933887951 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:14.937783191 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:15.164885235 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:15.168738381 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:15.174288038 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:15.177953056 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:15.510470586 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:15.514337308 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:15.510578720 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:15.514466291 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:15.390264795 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:15.390258149 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:15.393942073 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:15.393965621 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:15.390314444 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:15.393995371 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:15.569307332 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:15.573159213 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:15.573509949 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:15.577072388 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:15.500149569 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:15.578774095 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:15.503884030 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:15.500482625 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:15.411624581 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:15.582484880 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:15.504202057 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:15.500806227 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:15.415654514 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:15.411708018 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:15.504438991 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:15.415692349 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:15.500882205 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:15.504683064 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:15.514057390 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:15.517721806 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:15.514128226 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:15.518021134 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:15.473670337 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:15.473670441 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:15.478461088 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:15.478478517 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:15.878209600 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:15.882009940 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:15.894796521 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:15.898567937 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:15.960539427 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:15.964256214 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:15.937963878 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:15.941626757 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:16.168913366 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:16.172976290 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:16.178126275 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:16.181835412 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:16.514475215 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:16.518427629 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:16.514650677 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:16.518558667 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:16.394105715 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:16.394115401 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:16.397789754 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:16.397817147 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:16.394184713 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:16.397881820 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:16.573325678 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:16.577213975 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:16.577238722 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:16.580844204 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:16.504010903 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:16.582628869 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:16.507676579 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:16.504317635 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:16.415801761 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:16.586283712 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:16.507936613 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:16.419870533 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:16.415868964 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:16.504585837 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:16.419912563 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:16.508249142 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:16.504915892 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:16.508594338 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:16.517936435 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:16.521910095 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:16.518256551 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:16.521964331 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:16.478625444 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:16.478625384 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:16.483236102 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:16.483251422 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:16.882216616 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:16.886188642 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:16.898708819 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:16.902432088 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:16.964377051 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:16.968160312 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:16.941805853 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:16.945571006 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:17.173160058 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:17.177070689 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:17.182000952 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:17.185649476 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:17.518531331 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:17.522291351 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:17.518731776 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:17.522602694 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:17.397935802 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:17.577379066 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:17.507787360 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank12]:[W621 21:59:17.397992598 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:17.581223958 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:17.511578244 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:17.401639548 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:17.581001159 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:17.508097399 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:59:17.398048637 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:17.584510761 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:17.401669937 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:17.586429209 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:17.511912316 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:17.401685897 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:17.508393385 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:17.420071186 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:17.590133665 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:17.512015820 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:17.424178643 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:17.508754983 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:17.420109082 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:17.512371550 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:17.424301178 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:17.522162595 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank13]:[W621 21:59:17.483415828 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:17.526012567 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:17.522166415 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank10]:[W621 21:59:17.483415905 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:17.526056490 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:17.488078471 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:17.488097803 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:17.886385395 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:17.890210138 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:17.902565453 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:17.906278094 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:17.968320327 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:17.972046138 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:17.945758556 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:17.949578428 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:18.177210648 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:18.180991157 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:18.185787952 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:18.189420687 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:18.522417224 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:18.526401213 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:18.522776547 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:18.526634323 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:18.401786594 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:18.405515439 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:18.401849724 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:18.401915109 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:18.405555380 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:18.405572387 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:18.581400705 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:18.585345636 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:18.584676026 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:18.588525052 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:18.511689342 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:18.590282922 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:18.515410525 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:18.512161592 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:18.593968109 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:18.424331716 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:18.515794569 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:18.512079707 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:18.428384965 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:18.424610209 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:18.515881989 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:59:18.512531110 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:18.428607240 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:18.516155015 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:18.526226925 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:18.526257299 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:18.530162871 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:18.530197191 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:18.488274270 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:18.488274102 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:18.492931570 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:18.492947697 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:18.890413766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:18.894097598 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:18.906413082 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:18.910140200 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:18.972170789 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:18.975959227 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:18.949754627 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:18.953419560 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:19.181162237 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:19.185237950 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:19.189584350 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:19.193288088 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:19.526510744 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:19.530280266 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:19.526809541 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:19.530694269 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:19.405679561 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:19.405725685 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:19.409615374 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:19.405786014 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:19.409639759 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:19.409657056 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:19.585614648 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:19.589573487 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:19.588695821 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:19.592193489 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:19.515522125 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:19.594118489 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:19.519206758 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:19.515938280 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:19.597723212 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:19.519647515 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:19.428560881 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:19.516110977 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:19.432666408 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:19.428862900 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:19.519821368 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:59:19.516305085 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:19.432793837 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:19.519899097 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:19.530394410 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:19.530438830 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:19.534442526 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:19.534476829 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:19.493076346 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:19.493091288 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:19.497609349 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:19.497630011 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:19.894304465 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:19.898136737 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:19.910278216 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:19.914096349 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:19.976123978 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:19.979951076 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:19.953616464 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:19.957692829 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:20.185369157 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:20.189039679 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:20.193443134 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:20.197063742 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:20.530378062 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:20.534057591 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:20.530861963 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:20.534692796 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:20.409768962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:20.409810401 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:20.413731818 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:20.409870323 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:20.413760571 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:20.413781224 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:20.589742669 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:20.593502276 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:20.592346746 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:20.595859098 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:20.519313082 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:20.597854533 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:20.522969679 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:20.519781615 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:20.601518722 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:20.523414646 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:20.432813525 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:20.519960628 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:20.436865545 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:20.432962906 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:20.523600945 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:59:20.520061792 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:20.436910260 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:20.523656293 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:20.534677762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:20.538811099 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:20.534711536 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:20.538862800 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:20.497776467 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:20.497807635 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:20.502414331 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:20.502446099 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:20.898315338 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:20.901759872 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:20.914230285 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:20.918028464 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:20.980152625 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:20.983830824 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:20.957935136 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:20.961991755 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:21.189170960 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:21.192936151 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:21.197216580 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:21.201084068 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:21.534150772 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:21.537776496 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:21.534856640 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:21.538676262 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:21.413876218 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:21.417825136 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:21.413940157 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:21.413958469 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:21.417863346 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:21.417878624 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:21.593629837 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:21.597432326 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:21.596121211 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:21.600282756 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:21.523095033 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:21.601640238 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:21.526900285 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:21.523549671 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:21.605450171 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:21.527207350 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:21.437003419 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:21.523733254 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:21.440928251 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:21.437089280 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:21.527355349 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:59:21.523805738 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:21.440968928 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:21.527451368 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:21.539119763 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:21.542897957 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:21.539119872 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:21.542933945 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:21.502572130 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:21.502564949 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:21.507116827 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:21.507138336 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:21.901920490 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:21.905639848 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:21.918166619 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:21.922016054 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:21.984012586 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:21.987853931 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:21.962150205 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:21.966220914 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:22.193075291 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:22.196870117 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:22.201228759 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:22.204947472 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:22.537889781 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:22.541690188 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:22.538849190 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:22.542722346 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:22.417964525 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:22.421732160 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:22.418055928 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:22.418102211 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:22.421809128 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:22.421824251 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:22.597656933 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:22.601567363 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:22.600441985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:22.604429809 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:22.527012510 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:22.605582505 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:22.530609806 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:22.527343364 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:22.609318223 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:22.530925214 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:22.527486542 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:22.441074927 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:22.445038675 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:22.531054272 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:22.441162202 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:59:22.527731563 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:22.445082989 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:22.531482234 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:22.543113075 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:22.546969867 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:22.543229637 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:22.547015439 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:22.507455459 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:22.507455440 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:22.512021247 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:22.512036896 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:22.905821585 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:22.909498138 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:22.922152785 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:22.926015452 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:22.988040868 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:22.991890090 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:22.966492218 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:22.970709595 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:23.197001556 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:23.200956475 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:23.205098335 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:23.208751520 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:23.541836547 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:23.545762183 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:23.542877432 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:23.546701526 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:23.421897754 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:23.425831628 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:23.421993762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:23.422014488 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank25]:[W621 21:59:23.601740603 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:23.425865752 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:23.530710766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:59:23.425881382 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:23.605591812 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:23.445188488 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:59:23.604580791 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:23.534384453 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:23.531185082 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:23.608732572 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:23.448985518 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:23.445261279 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:23.609423815 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:23.534855698 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:23.449021886 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:23.531188329 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:23.613198668 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:23.534933195 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:23.531644671 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:23.535403404 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:23.547236503 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:23.551115012 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:23.547291366 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:23.551159718 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:23.512167783 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:23.512184315 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:23.516855620 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:23.516872145 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:23.909677655 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:23.913254486 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:23.926152773 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:23.930057362 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:23.992071898 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:23.996036207 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:23.970870524 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:23.974677570 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:24.201085231 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:24.204879873 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:24.208898938 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:24.212596636 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:24.545862769 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:24.549554541 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:24.546854977 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:24.550604499 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:24.426000786 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:24.429858675 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:24.425980039 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:24.426060376 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:24.429895929 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:24.429919201 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:24.605769373 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:24.609686822 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:24.608883284 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:24.612610559 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:24.534544489 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:24.613301401 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:59:24.538393817 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:24.534989520 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:24.616960754 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:24.538653233 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:24.449203807 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:24.535065869 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:24.453024887 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:24.449149251 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:24.538748376 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank17]:[W621 21:59:24.535570572 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:24.453061752 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:24.539356599 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:24.551326033 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:24.555006034 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:24.551344407 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:24.555311952 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:24.517007814 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:24.516991300 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:24.521795462 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:24.521795474 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:24.913436260 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:24.917028435 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:24.930191064 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:24.933962008 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:24.996211102 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:24.000200969 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:24.974870818 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:24.978962248 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:25.205029357 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:25.208880206 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:25.212758052 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:25.216480626 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:25.549663840 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:25.553381544 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:25.550759355 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:25.554508219 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:25.430051108 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:25.433981170 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:25.430052263 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:25.430104810 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:25.434014843 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:25.434035908 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:25.609863174 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:25.613803819 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:25.612768078 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:25.616462687 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:25.538560163 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:25.617082515 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:59:25.542374205 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:25.538897255 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:25.453171383 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:25.620874199 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:25.456985733 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:25.538828222 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:25.453211493 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:25.542439915 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:25.457019488 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:25.542461445 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:25.539524743 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:25.543258715 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:25.555220301 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:25.559034061 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:25.555507612 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:25.559073242 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:25.521943310 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:25.521943545 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:25.526610285 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:25.526612976 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:25.917180624 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:25.920738229 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:25.934096706 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:25.938054974 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:25.000367296 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:25.004230045 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:25.979162484 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:25.983407721 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:26.209030762 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:26.213021089 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:26.216616863 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:26.220413406 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:26.553479693 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:26.557060853 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:26.554662232 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:26.558339556 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:26.434148158 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:26.438109534 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:26.434172413 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:26.434219424 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:26.438144103 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:26.438163411 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:26.613971560 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:26.617850559 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:26.616616900 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:26.620265174 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:26.621005516 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:26.542607087 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:26.624674876 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:26.546298761 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:26.457129985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:26.542712290 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:26.457175414 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:59:26.542583321 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank14]:[W621 21:59:26.460951434 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:26.460981716 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:26.546363301 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:26.546390486 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:26.543409068 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:26.547192542 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:26.559266864 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:26.563060910 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:26.559246425 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:26.563459577 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:26.526759252 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:26.526759360 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:26.531382006 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:26.531381990 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:26.921028099 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:26.924800193 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:26.938189876 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:26.941839352 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:26.004398676 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:26.008223082 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:26.983601962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:26.987904883 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:27.213162257 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:27.217121026 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:27.220567494 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:27.224259467 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:27.557164661 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:27.560879388 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:27.558512821 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:27.562339889 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:27.438281309 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:27.438347929 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:27.442433146 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:27.442441648 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:27.438303839 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:27.442518510 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:27.618067985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:27.621929763 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:27.620409632 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:27.546441667 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:27.624120738 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:27.550270066 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:27.546514322 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank14]:[W621 21:59:27.461096088 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank30]:[W621 21:59:27.624806614 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:27.465047063 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:27.628583042 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:27.546526992 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank11]:[W621 21:59:27.461146598 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:27.550305467 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:27.550327030 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:27.465093407 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:27.547357527 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:27.551131876 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:27.563307454 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:27.567371700 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:27.563720358 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:27.567492223 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:27.531565282 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:27.531564985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:27.536189131 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:27.536191432 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:27.925058213 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:27.928733238 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:27.942004367 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:27.945863951 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:27.008364272 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:27.012212858 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:27.988100577 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:27.992052722 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:28.217260594 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:28.221181017 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:28.224407846 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:28.228030214 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:28.560994098 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:28.564600243 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:28.562517959 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:28.566221785 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:28.442573463 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:28.446589117 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:28.442636775 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:28.446644670 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:28.442667490 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:28.446767264 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:28.622101477 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:28.550393126 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:28.626114512 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank27]:[W621 21:59:28.624284803 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:28.554000245 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:28.550441160 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank27]:[W621 21:59:28.627904042 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:28.628780582 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:28.465175605 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank23]:[W621 21:59:28.550433992 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:28.468873806 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:28.632683493 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:28.554039403 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:28.465234985 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:28.554054291 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank17]:[W621 21:59:28.551294515 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank11]:[W621 21:59:28.468950631 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:28.554934574 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:28.567599023 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:28.567756027 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:28.571650004 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:28.571680764 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:28.536334133 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:28.536333797 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:28.540787913 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:28.540787585 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:28.928942663 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:28.932811990 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:28.946012592 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:28.950034835 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:28.012349915 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:28.016159978 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:28.992250833 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:28.996160139 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:29.221315928 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:29.225248229 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:29.228176336 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:29.231835212 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:29.564702401 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:29.568408219 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:29.566391546 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:29.570163262 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:29.446728465 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:29.450570450 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:29.446913336 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:29.626286881 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:59:29.446833934 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:29.630268493 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:29.450627931 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:29.628148198 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:29.554126948 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank15]:[W621 21:59:29.450647017 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:29.631676891 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:29.557851934 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:29.632851351 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:29.554199618 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:29.636703719 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:29.557946969 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:29.469004268 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:29.554268379 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:29.472729030 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:29.469090865 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:29.557990119 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +[rank17]:[W621 21:59:29.555192345 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:29.472761598 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:29.558897095 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:29.571936254 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:29.571877819 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:29.575800312 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:29.575821769 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:29.540934386 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:29.540934158 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:29.545608382 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:29.545625899 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:29.933079922 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:29.936952479 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:29.950202343 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:29.954012159 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:29.016337962 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:29.020400824 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:29.996329569 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:29.000326873 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:30.225458780 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:30.229436962 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:30.231986153 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:30.235735254 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:30.568512763 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:30.572194322 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:30.570319194 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:30.573988441 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:30.450815055 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:30.450726123 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:30.454568312 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:30.450789044 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:30.454586700 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:30.454606120 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:30.630437418 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:30.634391410 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:30.631826484 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:30.635464667 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:30.557975716 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:30.636835742 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:30.561753457 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank18]:[W621 21:59:30.558081201 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:30.640492794 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:30.558121081 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:30.561784044 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:30.561805820 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:30.559118737 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:30.562792090 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:30.472917515 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:30.476840323 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:30.472918557 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:30.476878648 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:30.576033210 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:30.576046999 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:30.579778838 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:30.579801891 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:30.545753926 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:30.545769386 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:30.550302692 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:30.550305690 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:30.937151298 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:30.941099888 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:30.954163032 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:30.957997346 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:30.020537029 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:30.024371142 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:30.000519156 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:31.004577293 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:31.229586012 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:31.233739084 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:31.235906102 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:31.239733420 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:31.572326854 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:31.576225872 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:31.574149385 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:31.577947158 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:31.454728919 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:31.458479078 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:31.454762053 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:31.454800028 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:31.458511875 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:31.458534278 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:31.634611142 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:31.638621247 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:31.635629502 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:31.639186447 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:31.561919577 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:31.640664637 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:31.565596902 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:31.561884469 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:31.644463981 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:31.565632184 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:31.561950972 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:31.565708075 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:31.562947569 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:31.566539762 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:31.477071320 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:31.481164219 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:31.477065402 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:31.481207191 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:31.579981445 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:31.583400829 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:31.580030659 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:31.583446981 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:31.550459949 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:31.550460018 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:31.554943131 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:31.554946064 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:31.941304766 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:31.944873068 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:31.958157764 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:31.962380309 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:31.024542343 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:31.028482061 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:32.004777451 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:32.008698420 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:32.233887075 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:32.237864026 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:32.239882882 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:32.243607552 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:32.576349056 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:32.580177127 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:32.578109939 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:32.581856793 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:32.458661778 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:32.458630125 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:32.462382661 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:32.458714867 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:32.462400459 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:32.462423010 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:32.638793312 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:32.642791834 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:32.639357815 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:32.642924841 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:32.565716213 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:32.644626012 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:32.569513004 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank20]:[W621 21:59:32.565799437 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +[rank30]:[W621 21:59:32.648402900 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:32.569582596 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:32.565878775 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:32.569681560 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:32.566693463 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:32.570336131 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:32.481361528 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:32.485266331 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:32.481361440 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:32.485301581 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:32.583587103 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:32.587539363 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:32.583654136 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:32.587575293 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:32.555086562 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:32.555086423 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:32.559860914 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:32.559862773 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:32.945077499 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:32.948674238 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:32.962551652 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:32.966437677 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:32.028636160 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:32.032469777 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:33.008897219 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:33.012811696 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:33.237995397 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:33.241806885 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:33.243727090 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:33.247364261 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:33.580281410 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:33.583928654 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:33.582014826 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:33.585782023 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank15]:[W621 21:59:33.462625393 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:33.462570267 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:33.466431525 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:33.462578361 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank9]:[W621 21:59:33.466449193 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:33.466468389 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:33.642961353 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:33.646811190 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:33.643088147 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:33.646921008 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:33.569672925 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank23]:[W621 21:59:33.573392971 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:33.648571645 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank20]:[W621 21:59:33.569789775 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:33.652348000 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:33.569849487 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank20]:[W621 21:59:33.573606509 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank18]:[W621 21:59:33.573624670 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank17]:[W621 21:59:33.570509377 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank17]:[W621 21:59:33.574263735 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:33.485444502 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:33.485491315 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:33.489395560 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:33.489418775 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:33.587798164 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:33.591622594 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:33.587752800 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:33.591662212 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:33.560155983 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:33.560156017 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:33.564763991 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:33.564763343 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank26]:[W621 21:59:33.948900156 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47068, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c8de7785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c8c7a5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c8c7a5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c8c7a5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c8c7a57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c884c509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c874c19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c8dfb35ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c8dfbc7850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank26]:[W621 21:59:33.952778896 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 26] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank29]:[W621 21:59:33.966578841 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47080, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e9553785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e93e65aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e93e65c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e93e65db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e93e657198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e8fb8509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e8eb819b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e9566b6ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e956748850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank29]:[W621 21:59:33.970295532 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 29] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank31]:[W621 21:59:33.032641374 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47092, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x150b8f1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150b7845aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150b7845c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150b7845db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150b78457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150b356509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150b25619b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x150b90583ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x150b90615850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank31]:[W621 21:59:33.036411339 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 31] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank8]:[W621 21:59:34.012986557 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-184]:47682, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e39bd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e38505aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e38505c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e38505db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e385057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e3422509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e332219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e39d091ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e39d123850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank8]:[W621 21:59:34.016758534 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 8] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank21]:[W621 21:59:34.241965494 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53918, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14e4883785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14e47125aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14e47125c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14e47125db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14e471257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14e42e4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14e487ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14e489462ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14e4894f4850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank21]:[W621 21:59:34.245818189 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 21] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank16]:[W621 21:59:34.247514366 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-877]:53932, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145833b785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14581ce5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14581ce5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14581ce5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14581ce57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1457da0509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1457ca019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145834ed2ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145834f64850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank16]:[W621 21:59:34.251367553 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 16] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank28]:[W621 21:59:34.584049481 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47024, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15064dd785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x150636c5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x150636c5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x150636c5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150636c57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1505f3e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15064d8f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15064ee7eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15064ef10850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank28]:[W621 21:59:34.587731114 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 28] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank24]:[W621 21:59:34.585920373 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=75, addr=[fs-mbz-gpu-880]:47032, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14fa4e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14fa3785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14fa3785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14fa3785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14fa37857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f9f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14fa4e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14fa4fa86ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14fa4fb18850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank24]:[W621 21:59:34.589552641 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 24] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank12]:[W621 21:59:34.466614714 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47714, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145b119785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145afac5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145afac5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145afac5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145afac57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145ab7e509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145aa7e19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145b12d47ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145b12dd9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:34.466615031 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47740, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1501bb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1501a485aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1501a485c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1501a485db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1501a4857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x150161a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x150151a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1501bc8aaac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1501bc93c850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank12]:[W621 21:59:34.470227998 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 12] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:34.466600886 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47698, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14c710d785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14c6fa05aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14c6fa05c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14c6fa05db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14c6fa057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14c6b72509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14c6a7219b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14c7120e3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14c712175850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank15]:[W621 21:59:34.470251628 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 15] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank9]:[W621 21:59:34.470271025 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 9] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank25]:[W621 21:59:34.647036096 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47042, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x148aacf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x148a95e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x148a95e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x148a95e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148a95e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148a530509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148aacaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x148aadf8eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148aae020850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank25]:[W621 21:59:34.650837663 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 25] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank27]:[W621 21:59:34.647193267 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47066, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14cb9ab785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14cb83e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14cb83e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank23]:[W621 21:59:34.573567643 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53952, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14f740f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14f729e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14f729e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14cb83e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14cb83e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14cb410509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14cb31019b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14cb9bf1eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x14f729e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14f729e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14f6e70509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14f740af1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14f74201dac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14cb9bfb0850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x14f7420af850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank27]:[W621 21:59:34.651023421 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 27] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank23]:[W621 21:59:34.577282552 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 23] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank30]:[W621 21:59:34.652520462 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-880]:47050, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14a2f5f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14a2df25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14a2df25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank18]:[W621 21:59:34.573820910 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53966, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153c6c3785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153c5525aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x153c5525c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14a2df25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14a2df257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14a29c4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14a28c419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14a2f72f7ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x153c5525db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153c55257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153c124509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153c6bef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153c6d47eac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14a2f7389850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x153c6d510850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank30]:[W621 21:59:34.656291981 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 30] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:34.573824123 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53974, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1491543785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14913d25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14913d25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14913d25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14913d257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1490fa4509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x149153ef1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x149155467ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1491554f9850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank18]:[W621 21:59:34.577561698 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 18] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank20]:[W621 21:59:34.577582415 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 20] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank11]:[W621 21:59:34.489557363 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47694, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x14ff2d1785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14ff1605aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14ff1605c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:59:34.574427563 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53936, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x151160f785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15114a25aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15114a25c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14ff1605db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x14ff16057198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x14fed32509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x14ff2ccf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x14ff2e1c0ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #3: + 0x5babb3e (0x15114a25db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x15114a257198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1511074509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1510f7419b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x151162316ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x14ff2e252850 in /lib/x86_64-linux-gnu/libc.so.6) + +frame #8: + 0x126850 (0x1511623a8850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank11]:[W621 21:59:34.493512470 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 11] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank14]:[W621 21:59:34.489597378 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47750, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x153eadf785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x153e96e5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +[rank17]:[W621 21:59:34.578107497 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 17] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +frame #2: + 0x5baa358 (0x153e96e5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x153e96e5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x153e96e57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x153e540509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x153eadaf1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x153eaf086ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x153eaf118850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank14]:[W621 21:59:34.493544435 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 14] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank19]:[W621 21:59:34.591839320 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53944, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1487905785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x14877985aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x14877985c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x14877985db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x148779857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x148736a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x148726a19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1487918b3ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x148791945850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank19]:[W621 21:59:34.595613889 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 19] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank22]:[W621 21:59:34.591881393 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-877]:53986, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x145eb67785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x145e9fa5aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x145e9fa5c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x145e9fa5db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x145e9fa57198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x145e5cc509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x145e4cc19b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x145eb7a9bac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x145eb7b2d850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank22]:[W621 21:59:34.595652928 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 22] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank10]:[W621 21:59:34.564901452 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47706, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x1462cb5785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x1462b445aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x1462b445c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x1462b445db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x1462b4457198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1462716509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x1462cb0f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x1462cc59fac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x1462cc631850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank13]:[W621 21:59:34.564917624 TCPStore.cpp:106] [c10d] sendBytes failed on SocketImpl(fd=95, addr=[fs-mbz-gpu-184]:47730, remote=[fs-mbz-gpu-124]:33087): Broken pipe +Exception raised from sendBytes at /pytorch/torch/csrc/distributed/c10d/Utils.hpp:653 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x15084e9785e8 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libc10.so) +frame #1: + 0x5ba8afe (0x15083785aafe in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #2: + 0x5baa358 (0x15083785c358 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #3: + 0x5babb3e (0x15083785db3e in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #4: c10d::TCPStore::check(std::vector, std::allocator >, std::allocator, std::allocator > > > const&) + 0x298 (0x150837857198 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so) +frame #5: c10d::ProcessGroupNCCL::heartbeatMonitor() + 0x379 (0x1507f4a509a9 in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so) +frame #6: + 0xd3b6d (0x15084e4f1b6d in /mnt/weka/home/hao.zhang/conda/miniconda/envs/junda-attnserver/bin/../lib/libstdc++.so.6) +frame #7: + 0x94ac3 (0x15084f993ac3 in /lib/x86_64-linux-gnu/libc.so.6) +frame #8: + 0x126850 (0x15084fa25850 in /lib/x86_64-linux-gnu/libc.so.6) + +[rank10]:[W621 21:59:34.569718677 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 10] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe +[rank13]:[W621 21:59:34.569742276 ProcessGroupNCCL.cpp:1659] [PG ID 0 PG GUID 0(default_pg) Rank 13] Failed to check the "should dump" flag on TCPStore, (maybe TCPStore server has shut down too early), with error: Broken pipe