[00:35] wxl: did we wind up with the broken busybox security update in those artful images? I'm not sure if that regression got fixed in artfu [00:38] slangasek: But I see nothing in artful-security for src:busybox? [01:39] slangasek: oh is that what it was? [01:50] i don't think that's it's slangasek. the security update was in bionic (1:1.27.2-1ubuntu4) but it looks like artful is still (it should be) 1:1.22.0-19ubuntu2 [01:56] wxl: What source package? [01:56] tsimonq2: busybox [01:56] the theory being our last debootstrap problem was due to busybox [01:58] wxl: ...src:busybox? O_o? [01:58] wxl: I don't see any updates post-release, that's the thing [01:58] https://launchpad.net/ubuntu/+source/busybox [01:59] look at the full changelogs for the current version and follow backward http://changelogs.ubuntu.com/changelogs/pool/main/b/busybox/busybox_1.27.2-2ubuntu3/changelog [02:00] Hmm [02:00] Why would that end up in Artful? [03:23] wxl: the security update was applied in all releases [03:24] AFAIK [03:24] hmm [03:24] slangasek: the changelogs don't reflect that and it's different symptoms [03:24] ok [03:24] I guess I'm just confused because it was fine for the 17.10 release, so something had to regress between then and now, and I see no busybox updates since the artful release... [03:25] yeah, looks like those weren't considered important enough security bugs to fix in stable releases [03:25] in that case, I have no clue what the cause is and would have to dig [03:30] is the failure reproducible with a debootstrap outside of d-i? [03:31] good question [03:32] never exactly done that before so i'll have to figure that out [03:32] @tsimonq2: you got a one up on me there? [03:34] so, lubuntu/artful/daily/pending/artful-alternate-amd64.list doesn't include libssl1.0.0 [03:35] slangasek: use-queued-livefs-builds> What you described is not the case, look at http://people.canonical.com/~ubuntu-archive/cd-build-logs/kubuntu/bionic/daily-live-20180104.log for example, it just gives up after a while and doesn't go forever. Am I wrong here? [03:35] I wonder if this is related to the recent cdimage changes related to seeds [03:35] wxl: Looking [03:36] slangasek: I doubt it, as it's a server thing too [03:37] tsimonq2: the daily jobs that started after launchpad was locked down only ended when I manually killed them [03:38] slangasek: Oh, makes sense. [04:31] ok, so in both the lubuntu and ubuntu-server failures, the packages that are not being found are a) not in the package pool where they should be, and b) packages that have been (security) SRUed. [04:37] Right [04:38] The weird thing is that build_all.sh claims to add libssl1.0.0 to the CD but then doesn't actually download it for some reason [04:38] (for Lubuntu) [04:40] Link from /srv/cdimage.ubuntu.com/ftp/pool/main/o/openssl/libssl1.0.0_1.0.2g-1ubuntu13.2_amd64.deb to /srv/cdimage.ubuntu.com/scratch/lubuntu/artful/daily/tmp/artful-amd64/CD1/pool/main/o/openssl/libssl1.0.0_1.0.2g-1ubuntu13.2_amd64.deb failed: No such file or directory [04:41] there we go [04:41] RIght [04:41] Theory: The script from debian-cd doesn't actually have artful-security in its sources.list. [04:41] http://people.canonical.com/~ubuntu-archive/cd-build-logs/lubuntu/artful/daily-20180105.log - see line 8408+ [04:41] 13.2 is newer than what was in artful release. And if it were looking for artful release version it would find it on the mirror. [04:42] the problem is we don't have an up-to-date Packages file for artful-security, it's looking for version 1.0.2g-1ubuntu13.2 instead of 1.0.2g-1ubuntu13.3 [04:42] likewise, curl (affecting ubuntu-server) has had *two* SRUs since artful release. [04:43] (here's what I was referring to before: https://paste.ubuntu.com/26344861/ ) [04:43] So it sounds right to me too. [04:43] good news, only lubuntu daily and ubuntu-server are affected by this failure [04:44] (i.e. nothing matching this in the logs for any of the other images, so it's not just a question of us not having seen the corner case yet) [04:45] ahaha [04:45] this is a bug of my own making [04:45] I configured rsync not to sync packages files for "releases we aren't updating" [04:52] tsimonq2, wxl: lubuntu daily, ubuntu-server images respinning now [04:52] slangasek: except it already failed :P [04:52] TypeError: execve() arg 3 contains a non-string value [04:53] well that's rude [04:53] tsimonq2: it failed, in your latest code [04:54] oh bah it's my fault again? XD [04:54] Alright... heh [04:54] run-tests is happy with it... reality is not [04:54] (was joking :P) [04:55] ((er, in my mind I was going to ask if you did ./run-tests)) [04:55] >_> [04:55] :D [04:55] hehehe [04:56] slangasek: Maybe I just have to surround GIT_TERMINAL_PROMPT=0 with quotes? [04:56] quite. I thought an earlier iteration did have quotes? where did they go? [04:57] Did it? Hm. [04:57] Not sure. [04:57] Either way, that's trivial, and you have commit access, you really want another MP? ;) [04:57] nope, I've already pushed the change [04:58] Cool cool [05:00] Gonna take off for tonight, o/ slangasek [05:00] ight [05:00] 'night, even [05:01] (what *is* it with gnome-shell losing the first characters after every window switch? :P) [05:05] -queuebot:#ubuntu-release- Builds: Lubuntu Alternate amd64 [Artful Dot One] has been updated (20180108.1) [05:06] -queuebot:#ubuntu-release- Builds: Lubuntu Alternate i386 [Artful Dot One] has been updated (20180108.1) [05:07] ls [05:09] slangasek: so the problem was in cdimage?? [05:09] yes [05:10] well that's a relief :) [05:11] this affected server because they're also git-based for seeds? [05:11] conveniently, it means no invalidation of the "point release" snapshot and no need to respin all images [05:11] wxl: it affected server because server also includes a package pool on the CD (in addition to a livefs) [05:11] ahhh [05:12] specifically, because those are the two images that included a package pool that referenced packages that had been SRUed more than once in artful - once before I made my 20/20 hindsight change, and once after [05:12] oh wow [05:13] -queuebot:#ubuntu-release- Builds: Ubuntu Server amd64 [Artful Dot One] has been updated (20180108.1) [05:13] -queuebot:#ubuntu-release- Builds: Ubuntu Server i386 [Artful Dot One] has been updated (20180108.1) [07:20] flocculant, wxl the download urls should all be fine now on the tracker for the point release [08:10] jibel: thank you! === lan3y is now known as Laney [11:53] jibel: Out of curiosity, what did you do to add those? I added them manually for Lubuntu a day ago because I couldn't find a better way. [11:54] wxl: ftr, Lubuntu is the only one with a Git-based seed at the moment. === maclin1 is now known as maclin [14:55] -queuebot:#ubuntu-release- Unapproved: postfix (xenial-proposed/main) [3.1.0-3ubuntu0.2 => 3.1.0-3ubuntu0.3] (core) [16:06] -queuebot:#ubuntu-release- New binary: linux-signed [amd64] (artful-proposed/main) [4.13.0-24.28] (core, kernel) [16:14] -queuebot:#ubuntu-release- Unapproved: gnome-software (xenial-proposed/main) [3.20.5-0ubuntu0.16.04.7 => 3.20.5-0ubuntu0.16.04.8] (ubuntu-desktop) === smoser1 is now known as smoser [16:53] I see the kubuntu tests for artful dot one show oem test failures; are these regressions? are they blockers for kubuntu images to be re-releaseD? [17:38] slangasek: It's likely a regression, I thought OEM installs worked, but unless acheronuk has the time or wants to give more details, I can dig into it when I get home in ~ 4 hours. [17:39] s/worked/worked for the 17.10 release/ to be more verbose... [17:46] tsimonq2: there were also some 'pass' results reported for the same test [17:49] slangasek: Hm, alright. [17:56] jibel: thanks [18:14] slangasek: If I wanted to document how the conversion to Git seeds was done, where do you think the best place would be to do that? [18:25] tsimonq2: wiki.ubuntu.com, I would imagine [18:35] oem failed for artful final, so not a regressions as far as I know, it's failed for a few releases now. [18:36] I think cyphermox was going to look at it, but what with one thing and another it's still outstanding [19:04] slangasek: Right, but any specific parent page I should put it under? [19:04] no idea :) [19:04] Ok, thanks anyways. [19:05] * acheronuk wonders how DarinMiller got oem to pass! [19:17] that's why I checked on legacy ... [20:59] slangasek: wiki.ubuntu.com/Germinate/ConvertingToGit - could I please get a second set of eyes on it when you have a moment?e [21:07] -queuebot:#ubuntu-release- Unapproved: linux-gcp (xenial-proposed/universe) [4.13.0-1003.6 => 4.13.0-1006.9] (kernel) [21:07] -queuebot:#ubuntu-release- Unapproved: linux-meta-gcp (xenial-proposed/universe) [4.13.0.1003.5 => 4.13.0.1006.8] (kernel) [21:09] -queuebot:#ubuntu-release- Unapproved: rejected linux-gcp [source] (xenial-proposed) [4.13.0-1006.9] [21:09] -queuebot:#ubuntu-release- Unapproved: rejected linux-meta-gcp [source] (xenial-proposed) [4.13.0.1006.8] === mwhudson_ is now known as mwhudson [23:22] Hm. slangasek, do you think Git seed support is needed in derive-distribution as well? Is that actively used for stuff anymore? (And if so, what for?) [23:22] (in lp:ubuntu-archive-tools) [23:23] * tsimonq2 looks around for docs on that tool [23:25] tsimonq2: I don't know; based on the timeline I think it was probably used for ubuntu-rtm (phone releases). You might double-check with cjwatson [23:26] Alright.