=== juergh_ is now known as juergh === hjensas is now known as hjensas|afk === cpaelzer__ is now known as cpaelzer === hjensas|afk is now known as hjensas [12:35] hi guys, just a quick question: how does cloud init translate "set-passwords" in cloud_config_module to "cc_set_passwords" ? Can't figure this one out [12:47] Well, I found out myself : https://github.com/canonical/cloud-init/blob/7c88f96245535aacfac9b0b04d9c0d82ac2067f4/cloudinit/config/__init__.py [13:12] blackboxsw, thansk for the review :) I will start working on the proposed changes today [13:36] oznt: Was that your entire question, or is there anything else we can help you out with? [13:38] Odd_Bloke, that was the question [14:18] OK, great. :) [17:06] falcojr: I've pushed the requested changes to both https://github.com/canonical/cloud-init/pull/354 and https://github.com/canonical/cloud-init/pull/349 :) [17:31] cool, I'll re-review now [17:32] I've technically already approved on the second one, you just need another approval from a committer [17:33] hello all - hope you are safe and well! just a quick question. when utilizing cloud-init to create templates, i have noticed that the startup script for network configuration is severely delayed. i have cloud-init to pull a DHCP lease, and this works without issue for directly standing up VM from an ISO. i am utilizing proxmox as the HV. [17:33] is there something i am missing to make this DHCP lease request, quicker... [17:42] blackboxsw: Can you add https://github.com/canonical/cloud-init/pull/349 to your list of PRs to double-check as core committer, please? [17:44] We also didn't assign reviews today, and I'm first so I've taken #358. [17:59] Odd_Bloke: yes I can/will [18:40] @Odd_Bloke, did you have a chance to take a look into the issue with the test_conftest having pylint error https://pastebin.ubuntu.com/p/QMkhNtShjW/ [19:30] AnhVoMSFT: I didn't, apologies! [19:30] Let me take a quick look now. [19:36] AnhVoMSFT: Hmm, could you give me a more recent log? I'm seeing differences in package versions, and it's hard to tell if that's just because there have been releases in the past ~week. [19:51] AnhVoMSFT: Hmm, actually, this is stranger than that. I can reproduce the error if I give the exact path (i.e. `.tox/pylint/bin/python -m pylint cloudinit/tests/test_conftest.py`), but not if I only specify the directory (i.e. `.tox/pylint/bin/python -m pylint cloudinit/tests/`). I'm also running Python 3.8 (instead of, I think, 3.5 in these logs). [19:59] AnhVoMSFT: And I also can't reproduce on 3.5, using tox. :/ [20:08] @Odd_Bloke yeah, I am having a hard time reproducing on my local machine - let me try with the .tox/pylint/bin/python. That should be the same version though [20:10] yes I can repro using that particular command [20:10] and I am on 3.5.2 on my xenial box (python3 --version) [20:10] the build server gets a new container every time it runs the test, so whatever is in .tox isn't cached between runs [20:46] AnhVoMSFT: OK, I think we're dealing with some sort of pylint bug here: I see different output for `.tox/pylint/bin/python -m pylint -j1 cloudinit.tests.test_conftest cloudinit.tests.test_util` and `.tox/pylint/bin/python -m pylint -j1 cloudinit.tests.test_util cloudinit.tests.test_conftest`. Can you confirm you see different behaviour also? [20:56] AnhVoMSFT: I'm filing a pylint issue now, will drop the link here once I have done so. [21:11] AnhVoMSFT: https://github.com/PyCQA/pylint/issues/3611 [21:20] yes @Odd_Bloke I can confirm I am seeing different behaviors between those two commands [21:21] how did the different ordering could come up in one build environment vs the others though? [21:21] Oh, is this because the container we were running the test had only one core ? [21:24] @Odd_Bloke: I guess the question now is: which one of them was the right one? (I know we want a consistent behavior, but which one is the "right" consistent behavior?) [21:39] AnhVoMSFT: So I think it's actually more likely that the container had many cores: I think the error is only reported if test_conftest is the _first_ module that a worker process tests (as I've just posted to the issue). [21:40] (The number of CPUs in the host is almost certainly what pylint will discover.) [21:41] AnhVoMSFT: So in this case, that test is specifically testing for behaviour when the wrong arguments are called (i.e. exactly what this pylint check is warning us about :p). So it's a valid pylint error, which we should be ignoring in this specific case. [21:41] However, I bet we also have a whole load of other instances of this error (or others which are similarly buggy). [21:52] lucasmoura: thanks one minor nit on https://github.com/canonical/cloud-init/pull/359 [21:52] then we can land that