/srv/irclogs.ubuntu.com/2016/01/05/#ubuntu-ci-eng.txt

=== chihchun_afk is now known as chihchun
robruyay05:10
michiwhat happened?05:10
michirobru: Do I need to rebuild?05:10
robrumichi: we had a buggy rollout, clobbered the creds which caused jos to hang. fixed it, but all the jobs had to be killed, which updated all the statuses unfortunately. this should settle down over the next 15 miutes..05:11
robrumichi: were you in the middle of a build just now?05:11
michiSo, don’t need to do anything?05:11
michiNo, the build had finished earlier.05:11
robrumichi: it's fine then, sorry for the spam05:11
michiWorks for me :)05:11
=== chihchun is now known as chihchun_afk
=== chihchun_afk is now known as chihchun
jibelbzoltan_, silo 56 fails to build on powerpc. Is it expected or not?09:14
jibelare you looking into it?09:15
bzoltan_jibel:  It is balck magic, sometimes it fails sometimes it does not. The failure if incorrect for sure as the library is there. We are looking at it, but it is not a critical issue in my view. the functionality of the UITKis more relevant09:17
Mirvlet me see if it happens every time09:32
Mirvbut yes they are a bit weird since earlier they were fixed but then it was found out it failed on s390x. then it was fixed even more and it built everywhere, and now once again it complains on another architecture (and only 32-bit version of it) where it used to work. some sort of compiler random order magic must be in play.09:33
=== chihchun is now known as chihchun_afk
=== _salem is now known as salem_
=== chihchun_afk is now known as chihchun
sil2100xavigarcia: piiing10:56
xavigarciasil2100: hi there10:58
sil2100xavigarcia: hey! I was looking at publishing of silo 51 and sadly there are some changelog issues there10:59
sil2100xavigarcia: I see that in https://code.launchpad.net/~xavi-garcia-mena/indicator-sound/restore-osd-notifications/+merge/281290 you modify the changelog - why is that?10:59
sil2100xavigarcia: we should never really touch existing history in most cases11:00
xavigarciasil2100: mmmm, I can't remember11:00
xavigarciasil2100: that was probably when I had to rollback trunk before Xmas11:00
sil2100xavigarcia: ok, could you maybe remove all the debian/changelog changes from your merges and rebuild indicator-sound? :)11:01
xavigarciasil2100: sure11:01
xavigarciasil2100: I will ping you when it is ready11:01
sil2100Since in the current state it's not really publishable, we even have some ugh, strange UNRELEASED versions in the changelog history11:01
sil2100xavigarcia: thanks!11:01
xavigarciasil2100: the changes in this branch.... https://code.launchpad.net/~xavi-garcia-mena/indicator-sound/re-add-integration-tests/+merge/281292 are needed, though11:04
xavigarciasil2100: is there a problem as they are?11:04
sil2100xavigarcia: hmmm, why can't those be put in the commit message instead?11:05
sil2100xavigarcia: so generally there should be no problem for having this debian/changelog modification there, but...11:06
xavigarciasil2100: I can't remember who it was, but I was asked to add those lines to the changelog11:07
sil2100xavigarcia: since the prerequisite branch has those invalid changelog modification/changes, this branch also carries some wrong bits in it (please notice that the indicator-sound (12.10.2+15.10.20151019-0ubuntu2) UNRELEASED; line is UNRELEASED)11:07
xavigarciasil2100: to reflect the new files and so on11:07
sil2100xavigarcia: so generally you can leave the debian/changelog modification in this branch if you make sure that the old changes from the invalid modifications are gone11:08
xavigarciasil2100: okay... so if I modify the prerequisite branch....should it be enough?11:08
sil2100xavigarcia: hmmm... you can try, depends if you have also merged the prerequisite branch to this branch as well11:08
sil2100xavigarcia: just make sure that before you rebuild, check the https://code.launchpad.net/~xavi-garcia-mena/indicator-sound/re-add-integration-tests/+merge/281292 merge diff if the previous changelog version visible is released to xenial - if it's UNRELEASED than something went wrong11:09
sil2100*then11:09
sil2100Only the newest changelog entry should be UNRELEASED11:10
xavigarciasil2100: Ok... will do that11:10
xavigarciasil2100: Okay, so I've removed the changes in changelog in the prerequisite branch11:18
xavigarciasil2100: but I still see UNRELEASED in https://code.launchpad.net/~xavi-garcia-mena/indicator-sound/re-add-integration-tests/+merge/28129211:18
sil2100xavigarcia: looking how this looks in reality11:34
sil2100xavigarcia: ok, I think it should be good, it's just a lag in LP it seems11:35
sil2100xavigarcia: try re-building :)11:36
xavigarciasil2100: well, now I have a conflict11:36
sil2100Oh, where?11:36
Mirvtvoss: davmor2: like https://requests.ci-train.ubuntu.com/#/ticket/793 says it needs a rebuild due to burned version number qtubuntu-media/vivid11:36
xavigarciasil2100: I did some extra changed11:36
xavigarciasil2100: changes11:36
xavigarciasil2100: I'm going to roll back to the one you saw11:37
tvossMirv, let me see11:37
tvossMirv, davmor2 should I just trigger a rebuild?11:37
davmor2tvoss: no idea I have nothing to do with that side, I guess it is because jhodapp uploaded a couple of media-hub fixes and might of bumped up the version number maybe11:38
Mirvtvoss: so it seems. at the same time the diff:s should get updated too. there was jhodapp's https://requests.ci-train.ubuntu.com/#/ticket/820 that landed yesterday11:39
Mirvdavmor2: missing diff:s on media-hub are not a problem, but a qtubuntu-media landing after the last build in that silo is11:40
tvossMirv, so I just trigger a rebuild, correct?11:45
Mirvtvoss: correct, but specify qtubuntu-media in the packages to build to make it even smaller re-build11:45
=== xavigarcia is now known as xavigarcia_lunch
=== marcusto_ is now known as marcustomlinson|
=== marcustomlinson| is now known as marcustomlinson_
=== marcustomlinson_ is now known as marcustomlinson
bzoltan_sil2100: Mirv: is there an information about what packages and what versions are in a certain  image version?12:30
sil2100bzoltan_: like, all packages in an image?12:31
sil2100bzoltan_: that would be in image manifests12:31
bzoltan_sil2100:  yes12:31
bzoltan_sil2100:  where can i find that?12:31
sil2100bzoltan_: http://people.canonical.com/~ubuntu-archive/cd-build-logs/ubuntu-touch/vivid/ - go here, find the rootfsid you're interested in, download the log and there you will find the link to the manifest (confusing, I know)12:33
sil2100bzoltan_: do you want a specific one? I can help you find it12:33
bzoltan_sil2100:  flo 88-89 diff i am interested in12:33
sil2100Ah, the diff12:33
sil2100bzoltan_: which channel?12:33
bzoltan_sil2100: ubuntu-touch/rc-proposed/ubuntu-pd12:34
sil2100bzoltan_: let me quickly grab the diff for you12:38
sil2100bzoltan_: hm, aparently there are no changes between 88 and 8912:39
sil2100bzoltan_: the manifests are - 88: https://launchpadlibrarian.net/232434413/livecd.ubuntu-pd.manifest and 89: https://launchpadlibrarian.net/232821809/livecd.ubuntu-pd.manifest12:39
bzoltan_sil2100: on flo? Wow... black magic - https://bugs.launchpad.net/ubuntu-rtm/+source/ubuntu-ui-toolkit/+bug/153093812:40
ubot5`Launchpad bug 1530938 in ubuntu-ui-toolkit (Ubuntu RTM) "header navigation broken on flo" [High,New]12:40
sil2100uh oh12:40
sil2100Yeah, but as you can see, manifests are the same and it actually makes sense since most people were out-of-office then12:41
sil2100The touch images also didn't have any changes12:41
bzoltan_sil2100: as expected12:42
sil2100Device and custom tarballs are intact12:42
sil2100So hm, maybe actually the whole 'reflashing' fixed the bug?12:42
bzoltan_sil2100:  how is that possible?12:43
oSoMoNtrainguards: can I please haz a silo for https://requests.ci-train.ubuntu.com/#/ticket/834 ?12:46
MirvoSoMoN: why wouldn't you assign it yourself, the safety limit shouldn't be hit?12:50
oSoMoNMirv, I tried, but got a message saying the number of free silos is too low12:56
oSoMoN(should I try again?)12:56
Mirvhmm12:56
Mirvrobru: the free silos limit shouldn't be as low as 52? yet it failed https://ci-train.ubuntu.com/job/prepare-silo/77/console12:57
Mirvrobru: oh, it says to me "No silos are available.". there's a glitch in the matrix.12:57
MirvoSoMoN: so no luck either from my side unfortunately. there are probably some ghosts haunting somewhere, meanwhile we'll need to wait for some silo to free up.12:58
Mirvwell, I can free one of mine12:58
oSoMoNok :(12:58
Mirvsince I'll follow it anyway12:58
MirvoSoMoN: no luck even that way. I freed up one silo and it still complains to me :(12:59
sil2100uuuuu13:00
Mirvhelp!13:00
Mirvsil2100: any idea, or should we just wait for robru?13:00
Mirvdebug run doesn't give any mor einsight13:00
=== xavigarcia_lunch is now known as xavigarcia
sil2100Let me take a look13:00
xavigarciasil2100: hi there13:00
xavigarciasil2100: the silo is built again13:00
xavigarciasil2100: could you please check if now looks better?13:01
sil2100xavigarcia: \o/ any changes besides the changelog?13:01
xavigarciasil2100: none13:01
xavigarciasil2100: just the changelog13:01
sil2100xavigarcia: will take a look in a moment, thanks!13:01
xavigarciasil2100: thanks13:02
Mirvtvoss: you didn't run the silo 22 rebuild, I did it now13:07
Mirvit seems jhodapp had tried to run the rebuild but the job execution had failed since force was not specified and the branch itself didn't have changes13:09
tvossMirv, ack13:10
tvossMirv, and thx :)13:10
Mirvdavmor2: the silo 058 is vivid only landing13:10
=== alan_g is now known as alan_g|lunch
sil2100Mirv: not sure if you got my previous message:13:14
sil2100Mirv: strange thing, the train doesn't even look for available silos from what I see - but the silo list is properly initialized13:14
davmor2Mirv: you say that but it doesn't work on vivid, which takes us back to what is it's point :)13:22
Mirvsil2100: no, that went to /dev/null apparently. ok then.13:30
Mirvdavmor2: yes, it doesn't change the end result :)13:30
Mirvdavmor2: tvoss: on another note, 022 rebuild now done and re-put into QA quque13:32
sil2100xavigarcia: hmmm13:38
sil2100xavigarcia: so I checked the silo again and hmmm... so, looks ok, but you removed the changelog modification from the re-add-integration-tests branch in the end?13:39
xavigarciasil2100: yeah13:41
xavigarciasil2100: it was getting mad13:41
sil2100xavigarcia: and the merge conflict on LP worries me, but if everything merges correctly when building then it's fine I guess - but I'm just wondering if you removed the big changelog entry with file mentions etc. by accident or planned13:42
xavigarciasil2100: it was planned... as there were lots con conflicts with the pre-requisite branches13:42
sil2100You could have just included all that in the commit message, the train should have just fetched all that and pasted it13:42
sil2100I mean, *should*, since in theory it shouldn't break anything13:42
xavigarciasil2100: you mean the LP commit message or when committing the code?13:43
xavigarciasil2100: I can add that in the comment in the MP13:43
sil2100LP merge commit message, the train always uses that for generating the changelog13:43
xavigarciasil2100: ok... will do that then13:43
xavigarciasil2100: one sec13:43
sil2100I mean, we can stick with what we want if the small changelog entry is enough, but I suppose it's a big branch13:43
sil2100So we might want to be more verbose on what appened13:43
sil2100*happened13:44
sil2100Fingers crossed that the train won't mangle it and break the formatting tho...13:44
xavigarciasil2100: OK... I've added the info in both, LP and the repository...13:51
xavigarciasil2100: rebuilding... will ping you when it's ready13:52
tvossMirv, great, thx13:54
=== alan_g|lunch is now known as alan_g
=== greyback__ is now known as greyback
jibelMirv, are we supposed to retest 22?14:57
sil2100robru: once you're up, could you take a look at the problem with assigning silos? I didn't look into much details, but from the first look the backend situation looks alright - from cyphermox-test I saw 50 silos assigned and in ~/.ci-train-silo-names I saw 61 silos defined15:03
sil2100robru: from the logs it looks like the train does not even really look for free silos, as I didn't see any logs from the "log_value_of.siloname('Checking if available')" bit15:04
=== sil2100 changed the topic of #ubuntu-ci-eng to: Train trouble? ping trainguards | CI problems? Please switch to Jenkins as a service: http://bit.ly/jenkins-docs | Train: http://bit.ly/1hGZsfS | QA Signoffs: http://bit.ly/1qMAKYd | Known Issues: issues with silo assignment
sil2100xavigarcia: hmmm, ok, rebuilding your silo, it seems we need to check the 'TAKE_WHOLE_COMMIT_MESSAGE' field15:06
sil2100Man, the rebuilds are killing me15:07
xavigarciasil2100: ok, thanks!15:11
davmor2morphis: any second15:15
morphisdavmor2: ?15:15
davmor2morphis: ^15:16
oSoMoNtrainguards: is the issue with silo assignments fixed?15:22
sil2100oSoMoN: no... robru will know what's up, he should be around shortly15:31
sil2100I didn't see anything obvious what could be wrong, and I don't have enough tools to debug that15:31
sil2100Since it's a closed environment15:31
xavigarciasil2100: Hi there, silo 51 is ready again15:43
sil2100xavigarcia: hey! Ok, so it's built, the changelog got a bit massacred (re-formatted) so it looks a bit ugly15:45
sil2100But...15:45
sil2100I suppose this is better15:45
sil2100https://ci-train.ubuntu.com/job/ubuntu-landing-051-1-build/lastBuild/artifact/indicator-sound_xenial_packaging_changes.diff15:46
xavigarciasil2100: so... can we land it now?15:48
sil2100Yeah, I'll do it now ;)15:51
morphissil2100: can you push something to a silo for me?15:51
sil2100morphis: sure15:51
sil2100:)15:51
morphisgreat15:51
Saviqfginther, hey, think https://launchpad.net/~jenkaas-hackers would be a good place for a jenkaas ML?15:57
fgintherSaviq, yes, I thought there was already one configured15:58
Saviqfginther, care to join #jenkaas on irc.c.c? :)15:59
fgintherSaviq, perhaps, are you aware that I now work for a different team (landscape)?15:59
dobeyeh, if that team is for people that hack on the thing, then it's not a good place fo ra support ML16:00
Saviqfginther, sure, doesn't mean you can't help here and there, your call :)16:00
Saviqfginther, still, you're listed as the owner of ~jenkaas-hackers ;D16:00
dobeya new team ~jenkaas-users would be better place for a user ML16:00
fgintherSaviq, that's true :-)16:01
Saviqdobey, don't think they're "hacking on the thing", at least the members list does not suggest that16:01
dobeySaviq: well, it's got a PPA and owns some branches16:02
dobeySaviq: so i'd think giving everyone direct commit access to those branches, who wants to join a ML to discuss user issues, is probably not a great plan :)16:03
Saviqdobey, sure, although it's the users who are the only hackers nowadays16:03
=== chihchun is now known as chihchun_afk
dobeyseems the wrong place for it, to me16:05
oSoMoNrobru, it seems the train is unable to assign silos to new requests, known issue?16:25
oSoMoN(I need a silo for https://requests.ci-train.ubuntu.com/#/ticket/834)16:25
sil2100Driving home, will be back in around 1h16:42
Saviqfginther, so people started requesting adding new folks (and whole teams) to ~jenkaas-hackers, maybe what dobey said makes sense - a lower-impact ~jenkaas-users team+ML could be a safer approach? I understand you might not want to maintain it... in which case a few of the initial users could be nominated as admins, but what happens to ~jenkaas-hackers?16:45
fgintherSaviq, I think that makes sense. The jenkaas-hackers group was setup because the ppas and code needed to be hosted somewhere. A users ML will help keep that separate16:50
fginthersorry for the slow response, in a meeting16:50
Saviqfginther, nw16:50
Saviqfginther, dobey, right, created a private ~jenkaas-users team+ML, owned by pspmteam17:04
davmor2tvoss, jhodapp: now fight to see who lands theirs first ;)17:46
jhodappdavmor2, haha18:03
jhodappkenvandine, are you up for landing something for me that requires a core dev? :)18:05
kenvandinejhodapp, sure, link?18:05
jhodappkenvandine, https://requests.ci-train.ubuntu.com/#/ticket/81218:05
kenvandinejhodapp, i'll look at it in a few18:05
jhodappkenvandine, sounds great thanks!18:05
=== alan_g is now known as alan_g|EOD
kenvandinejhodapp, interesting, so 5.4 provides 5.6 imports?18:07
kenvandinei guess there's no harm, just seems odd :)18:08
jhodappkenvandine, yes since our 5.4 is a backport of many things...it's to make sure we don't have to have any QML clients do version detection and remain compatible with what's in upstream18:08
kenvandineok18:08
jhodappkenvandine, also, Mirv looked this patch over already as well18:09
kenvandinejhodapp, what happens if an app imports 5.4 and uses Audio?18:09
jhodappkenvandine, they won't see the playlist stuff18:10
kenvandineok18:10
kenvandinejhodapp, done18:11
kenvandinethat was easy :)18:11
jhodappkenvandine, perfect, thanks!18:11
robruMirv: hey sorry for the hassle there, investigating now.18:31
=== jdstrand_ is now known as jdstrand
bfillerrobru: can you rebuild telephony-service on arm64 only in silo 3?19:38
robrubfiller: sure one sec19:38
robrubfiller: there's a bug in ticket creation at the moment, the assignment didn't work because your distribution field is wrong. please edit your ticket, correct the missing values, and then it'll work (I have a fix for this but it's not quite in production yet21:11
bfillerrobru: ok21:14
=== salem_ is now known as _salem
alesagerobru, do you happen to know if uitk folks (e.g. bzoltan_ ) care about powerpc builds? https://requests.ci-train.ubuntu.com/#/ticket/81823:24
robrualesage: if the train reports a failure in that arch then it's a regression on that arch, so either they will be forced to care or they will need a really good reason for regressing23:25
alesagerobru ack, ok, I'll pause this silo then and wait for a resolution (otherwise appeared passed)23:26
robrualesage: looks like it was rebuilt after being set ready for qa so just kick it back to them, yeah23:27
alesagerobru, ack thx23:27
robruYw23:27

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!