[06:18] xubuntu yak images failed to build this morning, not sure why other than it's Friday 13th ... [09:05] ^ Third time's the charm - nmqt now with a commit revert that breaks the ABI when built against nm ">= 1.2.0" (which is also in -proposed) [09:05] *reverted [09:41] mdeslaur, ^ synced over to unapproved queue for sru team review [10:02] flocculant: did you check the logs? [10:04] flocculant: looks like the same transient 503s from Launchpad that sil2100 was asking about in #ubuntu-devel re touch builds [10:04] flocculant: so I think my answer is the same, 503s can happen for various reasons and cdimage should probably be taught to backoff and retry [10:55] xnox: great, thanks [12:32] I did quickly look - then forgot and got on with morning stuff - thanks though :) [12:32] cjwatson: ^^ [14:52] pitti: see bug 1581535 please. [14:52] bug 1581535 in libnl3 (Ubuntu) "libnl upgrade breaks Network Manager" [Undecided,Confirmed] https://launchpad.net/bugs/1581535 [14:52] rharper: ^ [14:57] rbasak: meh, wasn't that supposed to be fixed in bug 1539634? [14:57] bug 1539634 in network-manager (Ubuntu Trusty) "network-manager crashes when using libnl-3-200-3.21.1-1ubuntu1" [High,Fix released] https://launchpad.net/bugs/1539634 [14:57] could somebody review that accountsservice upload ^ [14:57] pitti: was a Breaks correctly added? If not I'm wondering if some users have only a partial update. [14:57] it's a regression fix for a SRU that went to xenial-updates yesterday :-/ [14:58] wait [14:58] there might be an issue with the patch [14:59] grrr [14:59] pitti: I don't see one. [14:59] shrug, yes [14:59] fixing [15:01] rbasak: I guess that calls for another update then [15:01] sorry, I need to run in about 15 mins.. [15:01] I asked on the bug whether the reporters have the current NM [15:02] i. e. whether this is something different than 1539634 or just the missing breaks [15:03] pitti: if it is the missing Breaks, then I'm not sure another update will necessarily help. It might, but why are users picking up one update but not the other? [15:04] rbasak: not for the ones who already partially upgraded, but for new upgrades of course [15:04] we could also pull 3.2.21-1ubuntu1 from trusty-updates in the meantime [15:04] and then upload -1ubuntu2 with the breaks: [15:07] bah [15:07] ^ right one this time [15:07] pitti: that sounds reasonable. But I'm not confident I understand what's going on without a report that contains version numbers :-/ [15:07] Especially as it seems odd that some users would get a libnl3 update but not a network-manager update. [15:10] rbasak, rharper: I pulled the update from trusty-updates for now and followed up to both bugs [15:11] pitti, can you review accountsservice ^? sorry for the nag but it's a xenial-updates regression from a SRU copied yesterday :-/ [15:11] pitti: thank you for your help. [15:13] seb128: accepted, xenial task added; please upload to yakkety too, and nag some US archive admins (like bdmurray) to sru-release this in a few hours after build/test [15:13] * pitti needs to run, sorry [15:14] pitti, already uploaded to yakkety [15:14] pitti, thanks! [15:14] ah good, thanks [15:14] seb128: yakkety FTBFS [15:15] pitti, fixed already, that was my "ups/reject/reupload" dance [15:15] mismatched tag at line 15, column 27, byte 562 at /usr/lib/x86_64-linux-gnu/perl5/5.22/XML/Parser.pm line 187. [15:15] pitti, https://launchpad.net/ubuntu/+source/accountsservice/0.6.40-2ubuntu13.1 [15:15] Makefile:863: recipe for target 'org.freedesktop.accounts.policy' failed [15:15] seb128: ah, ok [15:16] * pitti waves [15:25] pitti, have a good w.e! === pete-woo_ is now known as pete-woods [16:11] bdmurray, some other SRU team member, can you look at https://launchpad.net/ubuntu/+source/accountsservice/0.6.40-2ubuntu11.1 ? see backlog [16:11] it's a xenial-updates regression [16:13] seb128: looking [16:13] bdmurray, thanks [16:16] seb128: and you want it released to -updates today? Is it urgent for some reason? [16:16] bdmurray, it's a regression from a previous update [16:17] if not can be block/delete the previous upload? [16:17] but it impacts some side feature [16:17] so no, not critical to get fixed today [16:17] the current accountsservice SRU is at 50%, so we could manually set it to 0 [16:25] bdmurray, your call, I don't know what are the standard rules for regressions, pitti seemed to suggest earlier than the regression fix should be moved over [16:25] but as said it's not a critical bug [16:25] so probably fine to wait next week [16:26] Given that 0.6.40-2ubuntu11 introduced a regression, I'd prefer to be cautious and release 11.1 on Monday when people are around to fix any other issues. [16:26] k [17:37] infinity: Could you halt the phasing of accountsservice due to a regression? 'change-override -z 0 -s xenial-updates -S accountsservice' should do the job [17:38] bdmurray: Iz done. [17:39] thanks [19:23] infinity: the change-override setting might have conflicted with the phased-updater. Could you run it again? [19:23] bdmurray: Did we delete all the arch:all binaries? :) [19:24] Oh, or you just went from 50 to 60 instead of 50 to 0. [19:26] right the phased-updater may have been running at the same time [19:26] 2016-05-13 18:26:03,150 - INFO - Incremented p-u-p for xenial-updates accountsservice from 50% to 60% [19:26] bdmurray: Yeah, will fix. [21:09] infinity: Have you seen bug 1581535? I'm wondering if there is more we should / could do. [21:09] bug 1581535 in libnl3 (Ubuntu Trusty) "libnl upgrade breaks Network Manager" [Critical,Confirmed] https://launchpad.net/bugs/1581535 [21:13] bdmurray: I reverted, and phased the revert to 100. [21:13] bdmurray: There's very little else we can do except cry. [21:14] infinity: Okay, I'll see if I can figure out how to do that. [21:15] bdmurray: Note that -1ubuntu1.1 in updates now is the same as -1 (ie: a full revert), -1ubuntu3 in proposed is the same as -1ubuntu1, but with the versioned Breaks that should have been there, and fixed symbols. [21:15] bdmurray: Figure out how to do.. What? Cry? [21:15] infinity: Yes, cry. [21:17] bdmurray: I'm mildly annoyed that the people who were aware of this bug in -proposed didn't think about the partial upgrade case. [21:17] bdmurray: But what's done is done. :( [21:18] I guess I could comment on the bug regarding the new libnl uploads [21:18] bdmurray: I did, but feel free to be more verbose. [21:20] bdmurray: The other workaround here to be mildly helpful might be to phase n-m to 100%. [21:20] bdmurray: So people who did get half upgraded might have a hope of being fully upgraded before their network explodes. [21:21] bdmurray: Gah. In fact, your scripts dropped it to 0%, which is exactly the opposite of helpful. [21:22] infinity: So fully phasing sounds like a good idea then. [21:22] bdmurray: Yeah, doing now. [21:23] bdmurray: AIUI, new n-m and old libnl are fine together, but the inverse explodes. [21:23] bdmurray: So we can fully phase n-m, hope that fixes some people by accident, then re-evaluate libnl on Monday. [21:24] infinity: I tested new n-m and old libnl in a VM and it seemed okay. The applet displayed and I could get to web sites. [21:24] bdmurray: Right, that's what I'd expect. [21:26] Very unfortunate to break people's networks, though, when that's the only way they can get fixes. :/ [21:26] I foresee a lot of people phoning their nerd friends this weekend. [21:27] bdmurray: Did you see the guy in one of the bug logs whose solution was "I bought a DVD player and reinstalled Ubuntu"? [21:27] A for effort, but ouch. [21:27] Surely, a DVD drive not player? [21:28] He said player, I assume drive. :P [21:28] Unless he just means he was so frustrated that he went on a manic electronics purchasing spree. [21:29] could have been a software dvd player [21:30] Anyhow, here's hoping the intersection of people who half upgraded and people who rebooted already before I pushed fixes and people who have no idea how to fix the result is reasonably small. :/ [21:30] Cause we're out of ways to help them remotely.