[05:37] xnox: i'm confused by openssl SECLEVEL patches in ubuntu https://launchpad.net/ubuntu/+source/openssl/1.1.1f-1ubuntu1 [05:38] specifically i don't understand why you: Revert "Enable system default config to enforce TLS1.2 as a minimum" & "Increase default security level from 1 to 2". [05:38] but then later on re-apply the same changes [05:39] it seems to me that you end up exactly where you started, with default SECLEVEL=2 and all the problems that causes [05:41] also it seems that ubuntu openssl is even more strict than the version in buster, and i don't understand why [05:42] and i'm not sure if it is even intentional [06:00] hmm ignore that last bit, i reproduced it with debian 10 as well [06:47] ddstreet: the rmq changes in groovy have been merged and uploaded to debian unstable so a re-sync should be possible [07:52] juliank, hi, I reworked by branch for the maas transitional package based on your input. I created a new MP since we want this in focal now, as maas is gone from the archive in groovy. Could you please take a look when you have time: https://code.launchpad.net/~ack/ubuntu/+source/maas/+git/maas/+merge/384013 ? [07:53] juliank, I linked the SRU bug as well, please let me know if that needs tweaking as I don't know the process well [08:20] ackk: ack === pieq_ is now known as pieq [09:06] Could anyone help me to upload https://bugs.launchpad.net/bugs/1875797? [09:06] Launchpad bug 1875797 in OEM Priority Project "[MIR] oem-somerville-melisa-meta" [High,In progress] [10:45] jamespage i'm confused about rabbitmq-server.service Wants epmd@.socket...it's not a template socket (or service)...? [10:45] i see lp #1808766 but i still don't get what it's doing [10:45] Launchpad bug 1808766 in rabbitmq-server (Ubuntu) "fails to start" [Undecided,Fix released] https://launchpad.net/bugs/1808766 [10:45] or trying to do [10:53] sil2100, Laney, xnox, bdmurray, we have a problem with a xenial update: https://bugs.launchpad.net/ubuntu/+source/json-c/+bug/1878723 [10:53] Launchpad bug 1878723 in json-c (Ubuntu) "Kernel panic when used with upstart after 0.11-4ubuntu2.1 update" [Critical,Confirmed] [10:54] this is taking down xenial instances with unattended-upgrades enabled [10:59] paride: doh, but I can't really help pull the update [10:59] leosilva: ubuntu-archive ^- [10:59] ouch, that's bad [11:00] * apw looks at the panic removal request [11:01] paride, i assume this was not happening before the json-c update, so it is that we are looking to remove [11:02] apw, that's my understanding too [11:03] if it's that bad an emergency chmod might be warranted - thoughts? [11:04] Laney, i am unclear if that is significantly quicker to propogate other than our mirrors [11:05] It doesn't help with propagation, but lots of people use our mirrors [11:05] It'll help people who use the Canoncial mirrors [11:05] cjwatson, and we would not remove it in that case ? [11:05] https://wiki.canonical.com/UbuntuEngineering/DealingWithCrisis#Limitation [11:05] You would, but this would mitigate immediately [11:05] ^- that [11:05] yes, that wiki page [11:06] That procedure has been tried and updated recently [11:09] At least it doesn't break unattended-upgrades, so as long as we get the new update out before people reboot we're not too bad. [11:09] hi, what's the status of this right now? [11:10] So it seems of high priority to push a revert with greater version. [11:10] ok removed; and the previous version is now copied back in [11:10] apw: Thanks, as long as we know that isn't the end of the relatively urgent story (as wgrant says) [11:10] cjwatson, yep we know that thanks [11:10] Is it just xenial? [11:11] i have only been told about xenial [11:11] Until we hear otherwise, let's prep reverts to -security of everything in that USN? [11:11] It was only issued to xenial and newer, and >= bionic doesn't have upstart [11:12] Ah yes, no pre-xenial indeed. [11:12] wgrant, how do we identify the USN [11:12] https://usn.ubuntu.com/4360-1/ [11:12] https://usn.ubuntu.com/4360-1/ [11:12] Ah, ESM [11:12] Ah yes [11:12] Conceivable it affects that too ... [11:12] That's annoying, technically does need security team then [11:12] Is anyone in a position to try? [11:13] Hi :) [11:13] Before it takes out half our DC or something [11:13] folks, I'm from AWS - joining to cooperate on the efforts [11:13] I have a VM I can sacrifice. [11:13] cjwatson, wgrant, I'm just going to build 0.11-4ubuntu2 as 0.11-4ubuntu2.2 in the security PPA and drop these changes. [11:14] to address a comment from wgrant above - we have actually heard from customers that this issue does indeed affect unattended-upgrades. [11:14] chrisccoulson: makes sense [11:14] rodarvus: I mean rather that it shouldn't break unattended-upgrades from upgrading to a fixed version, just that if someone reboots in between they're in trouble. [11:14] right. [11:14] I don't think it's worth trying to fix the existing patches on a friday afternoon [11:14] chrisccoulson: Please do. Ping me or Colin if there's any blocker at all. [11:14] Absolutely. [11:14] The vulnerability is way less bad than the breakage. [11:14] Straight revert. [11:14] chrisccoulson, if you need anything ping away [11:15] chrisccoulson: amazing. that sounds like the best way forward to me also. do you have any estimates on how long until the build complets and propagates to the security PPA? [11:20] ali1234: Ubuntu is more strict. I should publish my blog post about it. We enforce minimum protocol versions without configs. [11:20] Huh [11:20] Was it published to ESM? [11:20] I don't see it on my trusty systems [11:22] kenvandine, hi, it looks like that the Snap Store "_Permissions" string translations are not synced from Launchpad, see: https://bugs.launchpad.net/snap-store/+bug/1878672 [11:22] Launchpad bug 1878672 in snap-store ""_Permissions" string not translated" [Undecided,New] [11:30] wgrant: seems bionic is impacted too [11:30] ivoks: We'd like to know what impact people are seeing. [11:31] Since it can't break upstart there because upstart isn't used for boot. [11:31] wgrant: aws just reported that customers are complaining for bionic too now [11:31] wgrant: *cough* [11:31] xnox: ... oh? [11:32] xnox: What dirty secrets do you hide!? [11:32] We do not re-exec from upstart to systemd [11:33] xnox: You mean during an upgrade, pre-reboot? [11:33] Breaking that is deeply unfortunate, but not world-burningly critical. [11:33] Did we use upstart for user sessions in xenial? I forget [11:34] wgrant: revert libjson-c update obviously. As far as i can remember on package upgrades that upstart uses, we trigger stateful re-exec of init & all user-session upstart inits. [11:34] We have reports that this is affecting libjson-c3 version 0.12.1-1.3ubuntu0.1 in bionic too. See SF case # 00278097 [11:34] xnox: ... on bionic? [11:34] and clearly that's crashing at serialization, or deserialization [11:35] Oh or you mean it will break the system at upgrade on xenial, not at reboot? [11:35] philroche: "affecting" is very vague. [11:35] We'd really really really appreciate specifics. [11:36] wgrant: upstart is not published in bionic; but it can be still installed (xenial version). And i bet the maintainer hooks in libjson-c3 still try to restart it. [11:36] wgrant: thus i'd check that libjson-c3 stops trying to initrctl daemon-reexec or whatever it was called [11:36] xnox: Right, but breaking something that isn't /sbin/init is approximately one thousand times less critical than the situation on xenial, isn't it? [11:36] wgrant: it is vague :( Direct quote "Ubuntu will need to block the Bionic version as well, as it is also affected 0.12.1-1.3ubuntu0.1" [11:36] wgrant: if they have upstart still installed, it means they are forcing to still use it. [11:36] Ah [11:36] as pid 1 [11:37] wgrant: and upstart has support to be used inside chroots [11:37] philroche: The thing to sort out is whether that's just somebody observing that it's from the same USN, or whether it causes real issues (and if so what) [11:37] cjwatson: ack. I will get clarificaiton [11:38] rodarvus are you able to share details about how it affects bionic? [11:39] wgrant: telinit u || : is done by json-c in xenial; but not bionic. [11:39] Aha [11:39] wgrant: so it almost feels like "affects bionic" => where someone has xenial enabled in sources, and force installed upstart? [11:39] In which case they get to keep both pieces for the next two hours? [11:39] philroche cjwatson latest report from AWS is to ignore the 'affects bionic' comment. they are doing more testing before being sure about that. [11:39] Unless anyone strenuously disagrees. [11:40] ddstreet: ack [11:40] wgrant: should i be looking into fixing up upstart stateful re-exec here? [11:40] and or rebuilding upstart, to see the testsuite emit fire balls [11:42] ddstreet: we don't have access to customer instances, so unfortunately all we have is a number of customer tickets. I (personally) don't have any indication that is happened in bionic. [11:44] wgrant: rodarvus: ddstreet: at most, i'd expect dist-upgrades failing. [11:44] xenial => bionic, where one opted to stay on upstart in xenial. [11:44] xnox: right, thanks for confirming. [11:44] kenvandine, btw. there is another issue that was already fixed in upstream and it would be great if it could also be fixed in Snap Store (and deb gnome-software): https://bugs.launchpad.net/snap-store/+bug/1750818 [11:44] Launchpad bug 1750818 in gnome-software (Ubuntu) "Already installed deb packages not showing as installed when opened" [Low,Triaged] [11:44] (the fix is easy to backport) [11:46] So, now that reverts are in progress, and it seems likely that it is only properly fatal to xenial... [11:46] We should try to work out the various xenial scenarios [11:46] We have reports of OOMkills after upgrade and before reboot [11:46] We have reports of panics on reboot [11:46] And we have reports of OOMkills some time after reboot [11:47] Oh, I guess the second case may be the third, just soon after boot. [11:47] At least that means that booting and quickly upgrading should avert doom. The system isn't entirely unbootable. [11:53] Mirror triggering is disabled for the time being to avoid accidents [11:53] (from the master) [12:01] Update RE bionic "So bionic doesn't get the crash on init, and no kernel panic. But I am testing that the bug is still present in Bionic, just not triggered, as there is no upstart in Bionic. I'm looking to test the bug" [12:02] philroche: Thanks for confimring. === ijohnson is now known as ijohnson|lunch === ijohnson|lunch is now known as ijohnson === ben_r_ is now known as ben_r === ben_r_ is now known as ben_r === grumble is now known as rawr