[00:00] robru, sure... But it "remembers" today's build history.. But assigns the next job a number from yesterday I think : https://ci-train.ubuntu.com/job/ubuntu-landing-019-1-build/build [00:01] camako: yeah I can't explain that [00:01] camako: but your job is running, i think that's the best we can hope for [00:01] robru, yeah it seems to be working [00:01] thanks [00:01] camako: you're welcome! and sorry for the hassle! [00:46] camako: ^^ so your status being all "...1125 not in PPA", I'm assuming that's due to stale backups being restored, don't worry about that, I'm working on a fix for that, just leave it for now [00:47] robru, yea I was ignoring those === chihchun is now known as chihchun_afk [02:18] trainguards Hi all, it's been a while since I've used the train. After creating a request what's my next step? Do I assign it or is that used for something else? [02:24] traingruards, hey I am getting this error: 2015-12-01 21:24:58,594 WARNING Silo name list not found, run setup_citrain.py [02:24] traingruars, on silo 45, https://ci-train.ubuntu.com/job/ubuntu-landing-045-1-build/114/console [02:26] renatu: Hi, you may know :-) After requesting a silo, do I hit the Assign button or something else? [02:26] veebers, yes, and it was building before [02:27] renatu: sorry, I meant for my own silo. It's been a while since I've used it and things have changed [02:28] veebers, yes you need to hit the assign button [02:28] veebers, the silo should change to 'Ready to build" [02:28] renatu: ah I see, awesome thanks for the help :-) [02:33] renatu: the train imploded earlier today, how old is that log? Should work if you try again [02:33] robru, I am trying and nothing happens [02:35] renatu: what do you mean "nothing happens"? Doesn't look like you ran the job [02:35] robru, I click on build button and the screen refresh without any new job [02:36] renatu: one sec [02:40] renatu: it's something https://ci-train.ubuntu.com/job/ubuntu-landing-045-1-build/114/console [02:40] Brb, eating [02:49] robru: uhm, how are there two tickets assigned to silo 21? === chihchun_afk is now known as chihchun [02:49] i just tried to build https://requests.ci-train.ubuntu.com/#/ticket/724 but it is building location-service instead [02:50] which i see is already built in the ppa [02:58] dobey: the train experienced a catastrophic failure today and we restored from a backup. [02:59] dobey: was one of them landed already? [02:59] dobey: what ticket numbers? [03:01] i don't know about location-service [03:01] https://requests.ci-train.ubuntu.com/#/ticket/724 i created earlier today [03:01] just before the implosion [03:02] https://requests.ci-train.ubuntu.com/#/ticket/678 seems to be the one for location-service, and the MPs do say "Merged" on it [03:08] dobey: https://ci-train.ubuntu.com/job/ubuntu-landing-023-1-build/176/console you're building in 23 ho [03:08] now [03:08] ok [03:09] dobey: sorry for the hassle. [03:16] robru, the build started but got a very strange error [03:17] robru, https://requests.ci-train.ubuntu.com/#/ticket/722 [03:17] renatu: yeah that's because the backup is probably missing address-book-app. If you want to fix that quickly you can rebuild it, or wait some hours as I'm working on a permanent fix. [03:18] robru, ok I will rebuild it, thanks [03:18] renatu: ok, sorry === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [04:19] oh fun [06:21] here we go! [07:23] yay for working backups [07:33] Mirv: yay even more for pushing branches to lp so that backups aren't even necessary ;-) [07:33] robru: yeah, I read your e-mail, it sounds good [07:34] Mirv: I'm just going through and double checking that the silo states make sense and are consistent with reality. so far it seems good [07:36] robru: nice. thanks for your hard work. [07:36] it seems the Qt migration autopkgtests will take a while... this is 18h after the publishing: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#qtbase-opensource-src [07:37] Mirv: you're welcome. also sorry for breaking everything constantnly [07:37] Mirv: yes I've heard that autopkgtest infra is strained [07:37] robru: well when you break everything in every possible way, you eventually get a really tested and bulletproof system :) and the problems haven't like caused huge, longstanding problems. [07:38] Mirv: heh, thanks. once it stabilizes I think I need to take a month off ;-) [07:39] robru: yes, please eventually have a good, long break and we will also then ban you from IRC this time :) [07:39] lol [07:40] Mirv: ugh, silo 12 status reporting is broken, which means auto merge is also broken. looking into it. [07:45] robru: thanks. a working merge / marking as Landed would be nice. === chihchun_afk is now known as chihchun [07:46] Mirv: this error is very strange though as it seems to be an error in requests library itself. [07:49] Mirv: it seems to only happen with qtdoc-opensource-src. do you know anything unusal about that one? [07:51] robru: it's a package that had Ubuntu changes but now is a sync from Debian, that's the only thing special I can think of [07:51] Mirv: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-012/+sourcepub/5736443/+listing-archive-extra hummmm no signer!!! dun dun dun! [07:51] robru: right, if you do a copy-package from Debian to PPA or archive it doesn't have a signer indeed [07:52] Mirv: also no changes file? [07:52] (or so the LP says in the PPA) [07:52] robru: I guess it's because it's never uploaded to Ubuntu, it's a LP-made sync directly from Debian [07:53] Mirv: what's happening is that the train is trying to download the changes file from the ppa and from ubuntu and confirm that the contents of the changes file from both matches (ie, that the version in proposed is really our version from the ppa and not some other package with the same version number) [07:53] robru: there would have been many other ones, but the others were already auto-synced from Debian to -proposed and I made a "build1" of them [08:05] Mirv: ok, pushed a fix, we'll see how it goes in about 12 minutes. [08:19] robru: ok! [08:22] Mirv: so the train thinks your "pre-merge commit" is a new commit [08:24] Mirv: if possible I recommend deleting your commit, and then pushing https://code.launchpad.net/~ci-train-bot/qtubuntu-camera/qtubuntu-camera-ubuntu-xenial-landing-012 to lp:qtubuntu-camera instead. then the train will recognize it's own commit and report correctly that a rebuild is not needed [08:24] Mirv: as it stands, this 'new commit' warning will prevent the auto merge. [08:25] robru: ok, trying out [08:25] I tried to help Kaleo in working on his silo [08:26] robru: done [08:27] Mirv: ok, let's see if it works ;-) [08:27] no idea why I thought a manual merge would be a good idea. it was 1 min before leaving for team dinner.. [08:27] hehe [08:29] robru: it worked! [08:29] Mirv: blam [08:30] robru: that's nice actually, that train supports this pre-pushing while being able to track the migration still === chihchun is now known as chihchun_afk [08:31] Mirv: it didn't used to. before it would track tip commit id's and complain if there was any change. I fixed it to use 'bzr missing' and parse the output, so now it's much smarter about if commits are missing or not [08:32] that's smart indeed [08:33] Mirv: did silo 12 used to contain oxide and now it doesn't? [08:33] robru: correct. I moved the main packages to 059 for sil2100 to publish them. [08:33] ok [08:34] well, more exactly non-train main packages that needed to have a manual upload instead of no-change MP. [08:35] moved at the very last minute, worked well and thankfully sil2100 got his core dev rights. === chihchun_afk is now known as chihchun [08:36] Mirv: no worries, just found an old copy of the source tree from the last time oxide was diffed, just cleared it out [08:36] Mirv: https://ci-train.ubuntu.com/job/cyphermox-test/24/console how's that for a train full of silos? ;-) [08:42] robru: sweet! :) === robru changed the topic of #ubuntu-ci-eng to: Train trouble? ping trainguards | CI problems? ping cihelp | Train: http://bit.ly/1hGZsfS | QA Signoffs: http://bit.ly/1qMAKYd | Known Issues: train experienced a catastrophic failure but has been restored to normal operation. any problems, just rebuild your silo === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === _salem is now known as salem_ === xavigarcia is now known as xavigarcia_lunch [12:27] * sil2100 off to lunch === alan_g is now known as alan_g|lunch [13:35] kenvandine: Silo 14 approved === alan_g|lunch is now known as alan_g === chihchun is now known as chihchun_afk [14:27] Mirv, sil2100: one of you have time for some uploads? [14:27] morphis: I'm here o/ [14:28] awesome === barry` is now known as barry_ [14:40] Mirv: hi! For some reason whenever I build it, it tells me that it needs to be rebuilt: https://requests.ci-train.ubuntu.com/#/ticket/695 [15:04] mardy: funny. the only thing I see is that there's some disparency between your MP branch and the target - like trunk is at 11, the MP says unmerged commits start from 13. so maybe try doing a new MP based on trunk where you apply the diff in a single commit or something (just a guess based on seeing there's some history with the branch) === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [15:56] slangasek: hi again. can you poke qtpurchasing-opensource-src through the NEW queue in xenial? [16:00] whee netsplits [16:04] it'll unfortunately end up blocked in proposed though, i think, because of qt5.5 [16:33] dobey: if I were reviewing this package in a silo, I would reject it for [16:34] dobey: if I were reviewing this package in a silo, I would reject it for multiple packaging issues; but since it's in the NEW queue, I guess it's going through... [16:34] (and Mirv isn't around for me to poke) [16:36] slangasek: hmm, ok. yes, please bug mirv about that, as i'm sure the same issues need to be fixed in the vivid overlay version too [16:41] dobey: ok, he wasn't in this channel but his irc client is connected so I've braindumped to him and accepted the package [16:42] slangasek: ok, thanks [16:43] seems freenode is having lots of issues today [17:28] robru, sil2100: is there a spare silo I could take for a bit for testing some code changes? === alan_g is now known as alan_g|EOD [18:42] mterry: yeah there's like 15 free. You don't need to ask, generally ;-) [18:43] robru, yeah but there was that email a bit back where you said you folks were full up [18:43] didn't know if we were still hard up [18:43] mterry: what? That must have been over a week ago... [18:44] robru, yeah... :) [18:44] Point is, I'll make a silo [18:44] mterry: it says how many are in use on requests.c.u.c, just subtract from 60 total [18:44] So just 6 free [18:44] mterry: OK, you can make a request and assign it yourself, just let me know if there's any trouble [18:45] robru, that total would be interesting data to surface. like 54 / 60 Assigne [18:46] mterry: requests page unfortunately doesn't know how many total there are, unless I hard coded it. Eventually we'll move to a system of unlimited ppas, so it won't even matter [18:47] There were 15 free when I went to bed last night, people got greedy in my sleep ;-) [18:57] robru, xenial is having problems with qtdeclarative5-* packages. Do you know what is happening? [18:57] robru, full build log: https://launchpadlibrarian.net/228221248/buildlog_ubuntu-xenial-amd64.address-book-app_0.2%2B16.04.20151202.4-0ubuntu1_BUILDING.txt.gz [18:57] renatu: no, what? is there an error log? [18:57] robru, I ruined everything: https://requests.ci-train.ubuntu.com/#/ticket/737 [18:58] mterry: your merge is superceded: https://code.launchpad.net/~mterry/ubuntu-app-launch/warn-on-xapp/+merge/278497 [18:58] robru, ugh right [18:59] robru, oh [18:59] robru, did you see the "fail train" page? [18:59] Mirv: you still around? can you help renatu with this qt explosion? ^^ [18:59] mterry: I *made* the fail train page... what are you seeing now? [19:00] robru, well I was seeing that on that page (737) when I said I ruined everything. But now I don't see it. So I dunno [19:01] mterry: dunno, could be a transient network error or something. did it say an error code before it disappeared? [19:01] robru, no [19:01] then i dunno [19:02] robru, again: http://imgur.com/VuwWyBQ [19:03] Oh, at least queuebot tells me error [19:03] Hi, the ppc64el build on Xenial failed earlier today for https://requests.ci-train.ubuntu.com/#/ticket/726 and according to the build log, it looks like python3.5 seg faulted during setup. Any one know of this issue? [19:03] Buildlog: https://launchpadlibrarian.net/228197665/buildlog_ubuntu-xenial-ppc64el.libertine_0.99.6%2B16.04.20151202-0ubuntu1_BUILDING.txt.gz [19:03] mterry: uh well if you see {{'s like that it means javascript isn't running to fill out the template. are you running noscript or something? [19:03] robru, not intentionally :) [19:05] robru, huh. If I refresh the page, it doesn't go away. But if I press enter in the location bar, it does [19:06] mterry: try it in firefox. I think it's a problem on your end, it's fine for me [19:15] renatu: so long story short, there's a problem with qt and it'll take some time to fix. are you able to work on something else for a day or two? [19:16] ChrisTownsend: first I've heard of it [19:16] robru, can we land it? [19:16] robru: Hmm, I'll just do a rebuild and see what happens then. [19:17] robru, we will be able to land it, when it get approved, or we will need to wait for the fix? [19:17] ChrisTownsend: I'd report a python3.5 segfault directly to barry ;-) [19:17] renatu: well you can't land a package that won't build, no. [19:17] robru: Ok, I'll see if it happens again, and if so, poke him. [19:18] robru, ok [19:18] bfiller, ^^ [19:18] robru: Thanks [19:18] yw [19:19] renatu: what silo are you talking about [19:19] bfiller, all address-book silos [19:19] ChrisTownsend, robru: it's more likely of course that some extension is misbehaving [19:23] robru: we'll need to get that sorted rather quickly as it will potentially block a lot of silos [19:23] which I'm sure you already know [19:23] trainguards: for xenial FTBFS fixes on dual landed packages, is it required to go through the full QA process to land a trivial fix? [19:26] Hrrm, now all build are failing due to a different reason... [19:28] i'll presume not for now [19:35] heh [19:35] mterry: just a note, your qtmir-gles build will fail because you are trying to build qtmir 0.4.7 but qtmir-gles is still 0.4.6. If you're going to bump the upstream version number you have to do it in both branches [19:35] robru_hates_irc, ugh -- I was just trying to avoid the error about not having qtmir-gles [19:36] robru_hates_irc, I didn't want to bother making a real qtmir-gles branch [19:36] ChrisTownsend: next time please don't do a full source rebuild for a single-arch failure. we can retry single architectures [19:36] ChrisTownsend: random segfaults on ppc64el are likely a known (but fortunately quite rare) bit of guest memory corruption [19:36] mterry: no big deal, just s/0.4.6/0.4.7/ on the first line of the changelog in qtmir-gles [19:37] mterry: most of the -gles pain has been automated away [19:37] cjwatson: Ok, how do I just do a single arch rebuild? I don't see any obvious option for that. [19:37] ChrisTownsend: you have to ask me and I can do it [19:38] ChrisTownsend: anyone with direct upload access to the silo can do it, so anyone in ~ci-train-ppa-service [19:38] robru_hates_irc: Hmm, well, now I know [19:38] cjwatson: Ok, thanks [19:39] ChrisTownsend: there's an open bug about adding a train feature so people can retry ppa builds but it hasn't been a priority unfortunately. [19:40] cjwatson: robru_hates_irc: Since I did do the full source rebuild, any ideas why all xenial builds are failing? [19:40] err, xenial archive wonky at the moment? [19:40] ChrisTownsend: what silo #? i lost my scrollback [19:41] robru_hates_irc: 044 [19:41] https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-010/+packages has all build failures on xenial due to unable to install build deps [19:41] Yeah, maybe wonky archive. [19:42] i don't know when will be an appropiate time to have someone poke the retries though [19:42] haha yeah that was some amusement with archive admin [19:42] a certain person deleted qtbase-opensource-src from xenial-proposed [19:42] slangasek was just in #launchpad-ops asking for undeletion instructions, which I provided [19:42] cjwatson: is anybody working on a fix for that? timo just signed off... [19:42] ah [19:42] nice [19:42] so it should be back after the next publisher cycle [19:43] fun [19:43] cjwatson: how long are publisher cycles? [19:43] as long as they take [19:43] but typically 15 minutes ish [19:43] ah ok [19:43] yes; fixed on the archive side, and I'm addressing the removal with the responsible AA so that we don't end up in a revert war. [19:44] slangasek: I liked Mirv's plan to fix the contested issue in the next upload [19:44] rather than breaking the entire world [19:44] robru_hates_irc: Well, I'll need a xenial only rebuild for all archs for the silo when this is fixed:) [19:45] ChrisTownsend: will do [19:45] robru_hates_irc: Thank you! [19:45] yw [19:46] i guess a bunch of things will need that :) [20:07] ok, I retried that seconds too early, trying again [20:08] cjwatson: oh are you doing the retries? thanks [20:10] well, some [20:11] robru_hates_irc: I've done 10 and 44, but I haven't been keeping an eye on what else might need it [20:11] robru_hates_irc: also haven't touched the train (can't remember if that's needed nowadays), just silos [20:14] cjwatson: nah these days it'll notice things on it's own [20:14] thought so, good [20:23] renatu: ok, the issue with qt should be resolved, I've retried your builds [20:23] robru_hates_irc, thanks [20:23] robru_hates_irc, was quick than a day :D [20:23] renatu: yeah thanks to slangasek fixing it quickly, I was afraid it would take longer (qt is usually quite slow to move) [20:34] there we go === salem_ is now known as _salem [21:38] hmm [21:38] this channel has turned into a logfile [21:38] i guess i can't do that :-/ [21:42] robru: so having upload rights isn't enough to be able to publsih something? === ubot5` is now known as ubot5 [22:25] dobey: nope, it published. There's just a race condition between the status updater and the publisher, so it reverted to "successfully built" status temporarily before noticing the publish was successful [22:27] robru: ah, ok [22:27] ogra_: yeah i increased the granularity of the status reporting (now reporting per-arch statuses) and the bot went crazy pinging constantly. I'm not sure how to get it under control again [22:36] robru: fix the code to wait until all statuses are either failed or success, until reporting in channel? [22:36] (not sure how hard that would be, of course) [22:38] dobey: well the problem with that I'd that i want the first failure reported as early as possible so people can respond to it [22:39] I'd have to come up with a regex that ignores statuses that only have some combination of building/built but reports all failures [22:39] well, i mostly ignore the bot anyway [22:39] i should probably just actually /ignore it, so it doesn't even show up for me [22:39] dobey: right, it'd be worth paying attention to if it didn't spam constantly ;-) [22:40] well i just pay attention to the PPA page and the requests page [22:41] dobey: right but the advantage of the bot is that it's a push notification instead of having to reload the page constantly [22:42] Also i personally love the bot because i need to keep an eye on all failures, and investigate ones that are train problems [22:42] robru: implement actual push notifications support, and write a simple app for the phone? :) [22:43] dobey: ugh, push notifications ;-) I'd rather use cross platform tech like sending you emails [22:43] robru: there's a W3C proposal for push notifications :) [22:43] Ooooh [22:44] of course, i'm sure it's not implemented anywhere useful [22:45] I'll add support for w3c push and let the phone adapt to me ;-) === elopio_ is now known as elopio