/srv/irclogs.ubuntu.com/2013/11/29/#ubuntu-ci-eng.txt

robruogra_, rsalveti: time to kick an image build I think? cyphermox and I just landed the big mir transition. looking good on our end00:29
rsalvetirobru: cool, let me check if everything is already in00:30
robrursalveti, oh, and if you can ensure that python2.7 2.7.6-2ubuntu1 gets in, that would be super ;-)00:30
rsalvetisure00:31
rsalvetihm, python is still in proposed00:32
robrursalveti, yeah, just noticed that. it's absolutely critical for running AP tests though00:33
rsalvetiInvalidated by dependency00:33
rsalvetiNot considered00:33
rsalvetiDepends: python2.7 libffi (not considered)00:33
rsalvetiautopkgtest for git-annex 5.20131127.1: FAIL (Jenkins: public, private)00:34
rsalvetifor libffi00:34
robrursalveti, i'm not familiar with libffi00:34
rsalvetisync from debian, upload by doko00:35
Saviqcihelp, ps-nvidia-gt630 fails otto again http://s-jenkins.ubuntu-ci:8080/computer/ps-nvidia-gt630/builds :/02:03
Saviqagain the "could not start container" issue02:03
rsalveti=== Image r37 building ===04:06
rsalvetishould have latest mir/unity8/python2.7/etc04:06
* Mirv upgrades05:13
vilaSaviq: yeah, "again". Root cause https://bugs.launchpad.net/otto/+bug/1253198 , workaround https://code.launchpad.net/~vila/otto/stop-running-container/+merge/196594 waiting approval to be landed and deployed07:33
ubot5Ubuntu bug 1253198 in Otto "missing dependencies leaves the container running blocking further jobs" [Undecided,Confirmed]07:34
vilaSaviq: the container has been stopped manually in the mean time07:34
jibelvila, where is the code that install debs? I cannot find it in otto trunk07:42
jibelis it a custom branch of otto?07:42
vilajibel: I don't know of any otto custom branch, there are some uncommitted changes on most of the otto nodes but mainly for setting memory sizes07:45
jibelvila, it's in usr/local/bin/run-autopilot.sh, so not otto07:49
jibelin the script that calls autopilot07:50
jibelvila, I don't know where is comes from do you?07:50
vilajibel: first line of defense is to stop the otto container to unblock other jobs, that's the priority07:52
jibelvila, that script is broken07:52
jibelvila, there is a shebang with sh -eu07:52
jibelvila, but then it call apt with sh  -c07:52
ogra_hmm, these crashers in the ubnity8 tests on mako dont look good07:52
didrocksogra_: yep :/07:52
didrocksogra_: new mir, new crashers :p07:53
ogra_hud wouldnt worry me as much, but NM07:53
didrocksogra_: well, better to figure it out first07:53
ogra_right07:54
vilajibel: where do you see this script ? It's not on the otto node07:54
* ogra_ has a plumber in the house in 5min ... 07:54
ogra_not sure i'll make it in time for the call ... but i'll try my best07:54
didrocksogra_: no worry07:55
didrocksogra_: "swim" well ;)07:55
ogra_heh that was two weeks ago ...07:55
jibelvila, /var/lib/lxc/trusty-amd64-20131023-0529/run/delta/usr/local/bin/run-autopilot.sh07:55
ogra_i plastered a patch on the pipe ... thats what he has to fix properly today07:55
jibelon ps-radeon-835007:55
didrocksogra_: you used DEP3 headers I hope? ;)07:56
didrocksogra_: joke apart, I hope it won't cost you too much07:56
ogra_lol, nope, a tar patch07:56
didrocksinline patch! how ugly :p07:56
ogra_haha07:56
ogra_i dont care about the costs ... the pipes are 50 years old ... they all need replacement all across the 200m² over 3 floors07:57
vilajibel: can you review https://code.launchpad.net/~vila/otto/stop-running-container/+merge/196594 ?07:57
ogra_but i prefer that to happen in summer07:57
ogra_my house has 4 bathrooms ... lots of pipes :(07:58
vilajibel: and add https://launchpad.net/~canonical-ci-engineering to https://launchpad.net/~otto-dev ?07:58
didrocksogra_: yeah, summer will be better to shutdown the whole water system…07:58
vilajibel: making that team admin would be more in line with responsibilities too07:59
jibelvila, when do you want to run this script?08:07
vilajibel: This script is used as a post-build task in jenkins to catch containers left running. as mentioned in the MP08:08
vilajibel: and how to do that is described in the script itself08:09
jibelvila, but if you abort a job it won't be called, isn't it?08:10
vilajibel: and it has been tested on q-jenkins08:10
vilajibel: it is called even if the job is aborted08:10
vilajibel: it's explained in the script: This script is used as 'Post-build action' and as such is invoked after a08:11
vila16+# job is finished, including aborted.08:11
jibelvila, its annoying that it kills everything, can you restrict to the container that have been started by the test?08:13
jibelvila, as fginther suggested once, write a flag in the main job for example or limit to the containers following the naming convention used by the tests08:14
jibeldidrocks, or anyone, I'd need https://code.launchpad.net/~pitti/autopilot-gtk/fix-1254996/+merge/196703 in the distro, what's the procedure?08:17
jibelit is blocking installer tests08:17
didrocksjibel: just adding to the landing ask, I'm handling it08:18
vilajibel: further refinements can be done later (as explained in the MP)08:18
vilajibel: for now, I want to unblock Saviq and other users that face https://bugs.launchpad.net/otto/+bug/125319808:19
ubot5Ubuntu bug 1253198 in Ubuntu CI Services "missing dependencies leaves the container running blocking further jobs" [Undecided,New]08:19
jibelvila, it is not a refinement, it kills all running container without distinction08:19
vilajibel: It is indeed not a refinement, it is a workaround as explained in the MP and the script, refinements can come *later*08:20
jibelvila, for Saviq 's problem, fix the script run-autopilot.sh08:20
jibelused by medium tests08:20
jibelthis script is like saying "I've a bug on my machine, I don't know what it is, let press the power button"08:20
vilajibel: exactly08:21
vilajibel: and that's what I want08:21
jibelso you'd better fix the runner instead08:21
vilajibel: no, that workaround is a catch-up so the ci engine keeps running, other bugs will be fixed later08:21
vilacatch-all08:21
vilajibel: can I have that MP approved or https://launchpad.net/~canonical-ci-engineering added to https://launchpad.net/~otto-dev or should I just deployed my branch  on the otto nodes ?08:23
jibelvila, I disagree with the approach of killing all running container, please restrict to the containers used by the tests.08:24
vilajibel: fine, I'll deploy my branch then08:25
vilajibel: we can continue the discussion later08:26
Saviqvila, oh cool, nice that the problem is tracked down08:34
jibelvila, commented on the MP08:36
vilaSaviq: note that the root cause (wrong deps) won't be fixed by ci, someone has to look at them08:41
Saviqvila, of course08:41
Saviqvila, another one I've noticed from time to time bug #125622708:43
ubot5bug 1256227 in Ubuntu CI Services "Mako test runners sometimes fail with "write error"" [Undecided,New] https://launchpad.net/bugs/125622708:43
=== ev changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: ev | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: http://goo.gl/DStLaS
* ogra_ will be late to the meeting but i will come ...09:28
didrocksogra_: no worry!09:29
sil2100didrocks: can you ACK the packaging diff? Sortage mostly, as other changes were made by core devs ;p http://q-jenkins.ubuntu-ci:8080/view/cu2d/view/Head/view/QA/job/cu2d-qa-head-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_autopilot-gtk_1.4+14.04.20131128.1-0ubuntu1.diff09:49
Saviqev, hey, it seems makos are in a bad state this morning :/ https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/09:50
Saviqbug #125622709:50
ubot5bug 1256227 in Ubuntu CI Services "Mako test runners sometimes fail with "write error"" [Undecided,New] https://launchpad.net/bugs/125622709:50
ogra_Saviq, there are a ton of network manager crashes in the latest image smoketests too09:50
ogra_(no idea if that could be related)09:51
Saviqogra_, yeah, no idea either, don't get it what it complains about09:51
didrockssil2100: +109:52
ogra_popey, did you already try r37 on mako ?09:52
popeyam testing now ogra_09:52
ogra_the lockscreen issue on maguro is fixed09:53
popeynever saw that on mako09:53
ogra_yeah, that was driver specific09:53
ogra_didrocks, oh, btw, i would like to switch on cron after tonights meeting, would be good if everyone could think about the best times for the cron entry09:55
ogra_^^^^ rest of team too ^^^^09:55
didrocksogra_: hum, I would vary doing that while asac is sick/not around09:55
didrocksas he's the one disagreeing09:55
ogra_i would like to have some feedback data before he returns09:55
didrocksbut otherwise, yeah, we need to check with psivaa and plars in particular so that it matches their schedule09:55
ogra_we can still switch it off if he objects09:56
didrocksso that basically most of the tests are ran when they arrive09:56
didrocksthey can work on correcting/relaunching09:56
didrocks(so have some slack time)09:56
didrocksand then, we rekick the next one09:56
ogra_right, well, the tests have 8h to run09:56
didrockspsivaa: what would work for you? ^09:56
ogra_so i can set the schedule in a way that one image is just ready when they get up09:57
didrocksogra_: I would say, when psivaa starts his day, it would be better that the tests are already running for 3 hours09:57
ogra_or has run a chunk of automated tests already09:57
evSaviq: okay, adding to the queue. Thank you!09:57
didrocksyeah, exactly09:57
ogra_yeah09:57
popeyhttp://popey.com/~alan/phablet/device-2013-11-29-095719.png  anyone ever see that? where a key gets wedged on?09:57
popeynote it wasn't the last key I pressed, it stayed wedged even when pressing other keys09:57
psivaadidrocks: ogra_ i would like to have all the tests completed when i come in so that i could relaunch if there is any test that needs it09:58
didrockspsivaa: when are you starting UTC-wise?09:58
ogra_psivaa, when do you usually start09:58
ogra_heh, snap09:58
didrocks:p09:58
psivaaogra_: didrocks: it's UTC 9:0009:58
didrocksso half an hour before the meeting09:59
psivaaright09:59
ogra_so i should put the build say 4h before that ? does that suffice ?09:59
didrocksthis makes it 6 to start running the tests09:59
didrocksso 5 to build the image?09:59
didrocks5 + 8 = 1PM UTC09:59
psivaaogra_: that sounds good09:59
didrocksso tests done by 5 UTC10:00
didrockswhich is the other meeting time10:00
didrockshum10:00
didrocksbetter to get that done one hour before the other meeting10:00
didrocksso 410:00
didrockswe still have 3 hours for psivaa to fix things up in the morning10:00
ogra_ok, i'll schedule one build at 5am ... lets wait for plars input if such a schedule works for him too ...10:00
didrocksogra_: 4 would work better for the afternoon meeting10:01
ogra_we dont need to be strict on 8h or wanything10:01
didrocks4 -> 12 + 1 hour building + 3 hours of tests == 4 PM10:01
didrocks(so one hour before the evening meeting)10:01
ogra_that wont give plars much wiggle room10:02
ogra_i would start the second build ~ 2h earlier 13:00 UTC or so10:03
evheads up: Jenkins on s-jenkins just crashed on us10:03
popeyogra_: 37 is good here10:04
ogra_but lets wait for his input10:04
ogra_popey, no OSK for me10:04
didrockssil2100: once https://code.launchpad.net/~mardy/ubuntu-system-settings-online-accounts/disable-autopilot-tests/+merge/197152 is merged, can you add  ubuntu-system-settings-online-accounts to your bucket?10:04
popeyfine here10:04
ogra_popey, yeah, because you use a werid language/locale setting :P10:04
sil2100didrocks: sure! Keeping that on the radar ;)10:05
popeyogra_: CONFORM!10:05
psivaaogra_: didrocks: unity8 tests on both devices came good with 37. in transit to the dashboard10:05
didrockspsivaa: great!10:06
ogra_didrocks, so still no OSK on maguro (no surprise since nothing changed there)10:07
ogra_waking up the screen needs sometimes two power button presses ... but at least no tapping on screen anymore10:08
ogra_(still not 100% fixed though)10:08
didrocksogra_: ah, interesting, mind pinging greyback?10:08
ogra_about kbd or screen ?10:08
greybackno OSK on maguro! First I've heard of it.10:09
ogra_greyback, on non english locale10:09
ogra_greyback, bug 125599910:11
ubot5bug 1255999 in ubuntu-keyboard (Ubuntu) "[regression] if switched to non english local, the keyboard does not come up when an input field is focused" [High,Confirmed] https://launchpad.net/bugs/125599910:11
greybackogra_: ok thanks. I'll poke tmoenicke too10:11
ogra_he should be able to test a german setting i guess :)10:11
sil2100Mirv: https://code.launchpad.net/~sil2100/cupstream2distro-config/extra_for_qa/+merge/197157 could you take a lookie?10:14
ogra_psivaa, there was a saucy 102 build and i dont see any tests for it ... would be good to have some results within the next week (no hurry)10:14
psivaaogra_: ack, will add to the list10:15
Mirvsil2100: looks good10:15
sil2100Mirv: thank you!10:15
Mirvsil2100: one for you too https://code.launchpad.net/~timo-jyrinki/cupstream2distro-config/libavutil52/+merge/19715910:18
didrocksogra_: sorry, was chatting on something else, so no, on the power button10:20
didrocksgreyback:  ^10:20
ogra_didrocks, yeah, all sorted10:20
ogra_(see above)10:20
sil2100Mirv: oh, ok, not sure why but I'm somehow amused by this change ;p +110:20
Mirvsil2100: Friday? ;) I added another commit though, it seems it's installing both at the moment since I already did a really quick test run10:22
sil2100Mirv: ok, then I re-approve that10:22
sil2100Mirv: probably ;p10:23
Mirvthanks!10:23
Mirvlibav9 migration just completed today/yesterday, which was nice to notice10:24
popeybug 125626510:30
ubot5bug 1256265 in ubuntu-keyboard (Ubuntu) "Keyboard stuck down" [Undecided,New] https://launchpad.net/bugs/125626510:30
didrockspsivaa: so, can you flash latest image10:34
didrocks(locally)10:35
didrockstry to get previous glib version10:35
didrocksand see?10:35
didrocks(if the crashes still happens)10:35
psivaadidrocks: ok, i have a maguro and the crashes (network manager) is only happening on mako10:35
didrockspsivaa: argh, ok10:35
psivaadidrocks: network-manager somehow depends on libffi6 which is an upgraded package in 3710:37
didrockspsivaa: do you have access to device in the lab where you can test backing it out? (I'm testing locally)10:38
didrocksthat + retracing will really help I guess10:38
psivaadidrocks: yes, doing that. i have access to the device10:38
didrocksgreat, will confirm you shortly if I have the crash here (phone booting)10:38
didrockswith r3710:38
psivaaack10:38
didrockspsivaa: ok, so got the crash, reverting glib -> no crash FYI10:50
psivaadidrocks: ahh ok, thanks. i'll confirm10:51
didrockspsivaa: hands over a stacktrace once retraced10:51
asacseb128: ^^10:51
seb128asac, hey10:51
asacwill you backout given that glib has caused issues on touch and other parts?10:52
asac:)10:52
seb128asac, ssssssure10:52
asacor do you have a lead10:52
asaccool10:52
asacleading by example10:52
asachehe10:52
asacsorry... we need to get our proposed testing story sorted10:52
asacso we have a better safety net...10:52
psivaadidrocks: doesn't http://q-jenkins.ubuntu-ci:8080/job/trusty-touch-mako-smoke-ubuntu-rssreader-app-autopilot/39/artifact/clientlogs/_usr_sbin_NetworkManager.0.crash/*view*/ have the required trace information?10:53
seb128psivaa, it doesn't have debug symbols10:53
didrockspsivaa: do you know how to retrace a crash file?10:54
psivaadidrocks: i have not done in the past10:54
didrocksah, maybe we should pair someone with you10:55
didrocksMirv: mind giving a hand to psivaa? ^10:55
didrocksI think it's important that he learns about it10:55
didrocksyou have the -dbg files directly btw for glib10:55
Mirvdidrocks: sure, in a bit11:21
Mirvpsivaa: you'll need lp:daisy on host11:21
psivaaMirv: ok, i'll get it11:22
vilaSaviq: workaround deployed on s-jenkins, you may still encounter failing jobs but they should not block the following ones anymore11:25
Saviqvila, thanks11:25
sil2100didrocks: packaging ACK! http://q-jenkins.ubuntu-ci:8080/view/cu2d/view/Head/view/Misc./job/cu2d-misc-head-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_phablet-tools_1.0+14.04.20131129-0ubuntu1.diff ? :)11:25
didrockssil2100: +111:26
sil2100didrocks: thanks!11:29
sil2100hm11:29
sil2100Does anyone know why this is happening? http://q-jenkins.ubuntu-ci:8080/job/autopilot-trusty-daily_release/658/label=autopilot-nvidia/console11:29
sil2100I: No test left to run11:29
=== MacSlow is now known as MacSlow|lunch
psivaaseb128: didrocks: just an update.. still have not collected the crash debug information.. seeing issues like http://pastebin.ubuntu.com/6493858/12:36
didrockspsivaa: maybe, you can try retracing it directly on the phone?12:38
psivaadidrocks: let me try that12:38
didrockspsivaa: you apport-unpack the .crash file12:38
didrocksand then follow https://wiki.ubuntu.com/Backtrace on the CoreDump you will get12:38
rsalvetiogra_: do we have a bug for the weird nm bug you said already?12:38
didrocks(having the debug symbols12:39
ogra_rsalveti, not sure12:39
psivaadidrocks: ack, will try that12:39
ogra_i didnt file one (and dont have my mako on -proposed)12:39
rsalvetilet me flash the latest12:39
ogra_rsalveti, the dashboard has a ton of crash files though12:39
rsalvetioh =\12:39
ogra_http://ci.ubuntu.com/smokeng/trusty/touch/mako/37:20131129:20131129/5168/12:40
ogra_all tests with 2 or more crashes are likely to have the NM one12:40
seb128the nm one is a known issue which is being worked12:41
ogra_oh, cool12:41
rsalvetiseb128: right, but do we have a bug for it?12:42
seb128rsalveti, not sure, I've a fix in a ppa and I'm waiting for desrt to get online to validate it12:42
rsalvetialright12:42
seb128I confirmed the fix, I just want sanity check that I'm not doing something boggy to gobject12:42
Mirvdidrocks: I think the problem psivaa is seeing, based on the output he's getting, is that it's not a problem/crash that would create coredump12:47
didrocksMirv: I doubt there is no coredump, I can see one on the dashboard12:47
Mirvdidrocks: ok. it's what apport is telling him though, and I've had some of those with eg. python crashes. but it might be some different kind of problem.12:48
Mirvdidrocks: the method I gave psivaa I've used successfully with unity8 backtracing at lest12:48
Mirvleast12:48
seb128that report is working12:48
seb128I got the dump out this morning to look at it12:48
seb128that's not rocket science, just use apport tools (e.g unpack and gdb it)...12:49
vilasil2100: literally, I: No test left to run12:49
vilameans that all tests in the spool dir have run and there is no more to run12:49
MirvI'll try retracing too12:49
sil2100vila: but it results in an error, and the autopilot suite has more than 0 tests12:49
vilasil2100: this comes from run-autopilot.sh12:49
vilasil2100: that's not why it results in an error IIUC:12:50
sil2100I actually ran them on my desktop and there all tests ran12:50
seb128Mirv, don't bother, no need to be several doing it12:50
vila    if [ ! "$(ls -A $spooldir/)" ]; then12:50
vila        echo "I: No test left to run"12:50
vila        sudo shutdown -h +112:50
vila    fi12:50
seb128Mirv, that's dupping work for no good reason12:50
vilasil2100: that's the way it always finish the job12:50
Mirvseb128: well, I'm interested in seeing if I get similar error as psivaa, or not. my computer's CPU and network cycles are free :)12:51
Mirv(almost, at least)12:51
seb128Mirv, k, your call12:51
Mirvpsivaa got his .crash file from the device, so I'm checking if we get similar results on that downloadable one12:51
Mirvseb128: we were asked to pair work on this anyhow..12:51
Mirvpsivaa: yeah, on that jenkins one I got retrace, no problems12:52
=== MacSlow|lunch is now known as MacSlow
vilasil2100: i.e. there may be another previous error leading to *not* producing a test suite to run but I can't see why from that log12:53
vilasil2100: /var/local/autopilot/autopilot.log: Loading tests from: /usr/lib/python2.7/dist-packages kind of means autopilot has run and failed for an unknown reason (few lines above that 'No test left ro run' line)12:54
sil2100hmmm12:54
sil2100vila: thanks for the analysis!12:54
vilasil2100: my pleasure, learning with you here12:54
vilasil2100: in fact, I think autopilot has run indeed, yet a few lines before:12:55
vila/var/local/autopilot/autopilot.log: I: Running autopilot run autopilot -v -f xml -r -rd /var/local/autopilot//videos/ --record-options=--fps=6,--no-wm-check -o /var/local/autopilot//junit//autopilot.xml12:55
vilasil2100: so, was that job about running autopilot tests themselves ?12:58
vilasil2100: i.e. 'autopilot run autopilot' ?12:59
=== ev changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: http://goo.gl/DStLaS
psivaaogra_: i'm done with image 37.. please feel free to kick 38 :)13:14
=== psivaa is now known as psivaa-lunch
ogra_doing :)13:16
ogra_=== Image r38 building ===13:16
popey\o/13:38
Saviqev, hey, can you tell me how to subscribe to packages for errors.u.c? or where to read about e.u.c for that matter?13:57
seb128ev, define "subscribe"?14:00
seb128ups14:00
seb128Saviq, ^^14:00
seb128Saviq, like getting emails for issues?14:00
seb128Saviq, by default it lists packages you are subscribe to in launchpad14:01
Saviqseb128, how do I do that? on https://launchpad.net/ubuntu/+source/unity8 ?14:01
sil2100Damn, LP acts strange today..14:01
seb128Saviq, https://launchpad.net/ubuntu/+source/unity8/+subscribe14:02
seb128click the "I want to receive these notifications by e-mail."14:02
seb128Saviq, that doesn't give you emails from e.u.c though, but the default view is to list reports for the packages your user is subscribed to14:02
Saviqseb128, right, I was team-subscribed to that14:02
seb128Saviq, well, you can customize your url14:03
seb128Saviq, e.g https://errors.ubuntu.com/?release=Ubuntu%2013.10&user=unity-team&period=day14:03
Saviqseb128, right, thanks!14:03
didrocksok, exercising time14:03
seb128yw14:04
seb128Saviq, https://errors.ubuntu.com/?release=Ubuntu%2013.10&user=saviq&period=day14:04
Saviqseb128, what does it mean when a row is greyed out?14:04
Saviqretrace failed?14:04
seb128Saviq, I think it means it hasn't been seen with the current version of the package14:04
Saviqah14:04
Saviqnice one14:04
seb128so it might be fixed14:04
seb128crossed is when the bug is closed14:04
seb128red is when it's closed but the issue is still reported with the current version14:05
seb128would be nice to have a legend on the page14:05
seb128ev tried to make it "obvious enough by itself" but it's still confusing14:05
Saviqseb128, can it be greyed out if it's crossed, too? like https://errors.ubuntu.com/problem/f27a9186b6a83d502641e0fd89381bcd6209539514:06
Saviqseb128, it's fixed, not greyed out - but not found in current version14:06
Saviqprobably doesn't make sense to be greyed out if crossed14:06
seb128right, fixed is just crossed14:06
seb128it would turn red if it was seen in the current version14:07
Saviqseb128, aanyway, there anywhere I could read so that I don't bother you? :)14:07
seb128Saviq, try asking to mpt on #ubuntu-devel if there is a document explaining e.u.c or its design14:07
Saviqseb128, will do, thanks14:08
seb128yw14:08
=== psivaa-lunch is now known as psivaa
psivaaseb128: didrocks: Mirv has reported the bug for NetworkManager crash, bug #125629914:22
ubot5bug 1256299 in network-manager (Ubuntu) "NetworkManager crashed with signal 5 in g_object_finalize()" [Medium,New] https://launchpad.net/bugs/125629914:22
seb128psivaa, Mirv: you can close it, the fix has already been uploaded14:22
psivaaseb128: i'll wait for the fix to land and work before closing it14:23
seb128ok, your call14:23
seb128testing for the fix (that's the new glib just uploaded to trusty-proposed) is welcome14:23
=== Ursinha is now known as Ursinha-afk
didrockspsivaa: mind testing it from proposed?15:39
psivaadidrocks: the tests are running on the device. i was planning to test it once the tests are done15:41
psivaai mean the tests with r3815:41
didrockspsivaa: ok ;)15:41
dobeytransitioning stuff to daily-release is hard15:53
=== Ursinha-afk is now known as Ursinha
=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
Laneycan I get permissions to re-run autopkgtests?16:30
Laney(what happened to the matrix reloaded?)16:37
cjohnstonLaney: can you file a bug against ubuntu-ci-services-iteself with your request.. im not sure what may have changed with the move, but that way it can be followed up on when people from the US get back16:38
Laneycjohnston: ok, in the meantime can you re-run glib2.0 please?16:38
cjohnstonLaney: link please?16:38
Laneyhttp://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-glib2.0/21/16:38
cjohnstonLaney: I'm not familiar.. so I see 'matrix reloaded', on that page I see amd64 which is green and i386 which is red and checked...16:40
cjohnstonwhat about the downstream builds? do I check that?16:40
LaneyI don't know16:41
Laneythere used to be Matrix Reloaded to make the tests re-run16:41
Laneydon't ask me what it did, but I'm guessing it's a plugin which we don't have any more16:41
Laneyjibel: ^- any clue?16:41
cjohnstonLaney: the plugin still seems to exist (for me atleast)16:42
Laneyoh, you /do/ see it, sorry - misread16:42
cjohnstonLaney: I'm just not sure with the plugin, do  I just leave it at the default for what they set16:42
Laneyso tick both of them and press publish or whatever it is16:42
Laneythat should create a new entry to re-run the tests16:42
cjohnston"Matrix Reload downstream builds: "16:42
cjohnstonany idea on that tick box?16:42
Laneycan't remember16:42
LaneyISTR only changing the arches16:43
cjohnstonok, well its building16:43
Laneyneat, thanks16:44
LaneyI filed https://bugs.launchpad.net/ubuntu-ci-services-itself/+bug/1256372 for you16:44
ubot5Ubuntu bug 1256372 in Ubuntu CI Services "Can't retry failed autopkgtests any more" [Undecided,New]16:44
cjohnstonty16:45
=== Ursinha is now known as Ursinha-afk
cjohnstonI'm not sure if policies changed or what the story is, so I'm not comfortable with making changes :-)16:45
Laneyok, well I described my ideal16:46
cjohnstonty16:48
cjohnstoncyphermox: ping16:49
cyphermoxpong16:49
cjohnstoncyphermox: can you walk through updating the q-jenkins stack update stuff with me please?16:50
cyphermoxwhat do you mean?16:50
cjohnstonhttps://wiki.ubuntu.com/DailyRelease/FAQ#Adding.2BAC8-removing_components_to_a_stack16:50
cyphermoxok, what do you want to do?16:51
cjohnstoncyphermox: I'm not really sure what I'm doing, so just make sure I don't kill things?16:51
cyphermoxwell, that's what I mean what are you trying to do?16:51
cjohnstonapplying https://code.launchpad.net/~dobey/cupstream2distro-config/u1creds-inline-packaging/+merge/19723016:52
cyphermoxah!16:53
cyphermoxso since it's merged, just make sure you have the very latest revision16:53
cyphermoxthen you'll want to cd daily-release;  ./cu2d-update-stack -dUS ../stacks/head/webcred.cfg16:54
cjohnstoncyphermox: do I do this locally or on the remote machine?16:54
cyphermoxthat will update the jenkins config in jenkins, and not touch16:54
cyphermoxoh wait16:54
cyphermoxcu2d-update-stack -dU16:54
cyphermoxlocally16:55
cyphermoxyou should have a .cu2d.cred file on your system that has your credentials for jenkins to be able to update the config remotely16:55
cjohnstonhrm.. ok, I'll create oen16:55
cjohnstonone16:55
cyphermoxhmm16:55
cyphermoxok just a second16:55
cyphermoxcjohnston: so yeah, all you need to do is get lp:cupstream2distro-config, cd daily-release and then run16:59
cyphermox./cu2d-update-stack -dU ../stacks/head/webcred.cfg16:59
cyphermoxnote that this often requires you to be in the right teams to be able to update what trunk points to and things like that, so I'm not sure how well it's going to run17:00
cjohnstonack17:00
cjohnstonits runnin, i guess we shall see17:00
cjohnstonread only transport17:01
cyphermoxright17:01
cyphermoxwe also need https://code.launchpad.net/~dobey/ubuntuone-credentials/daily-release/+merge/197222 to get merged though17:01
cyphermoxlet me see if I can run it17:01
dobeycyphermox: no we don't17:02
dobeycyphermox: the purpose of the previous change was so that this can merge17:02
cyphermoxin that case this is not the same process17:03
dobeyeh? jenkins needs config that doesn't include pulling in the external package branch, to be able to run tests in this branch, because they conflict17:03
cyphermoxcjohnston: then you'll likely want fginther or alesage or some other person to fix up CI rather than daily-release17:03
cjohnstondunno.. this is what didrocks said was needed17:04
cyphermoxI expect there was a misunderstanding17:04
cyphermoxpackaging_branch: may be used in more than one place17:04
cyphermoxthe cu2d-update-stack stuff is really mostly going to change q-jenkins for the cu2d jobs, not the merge proposal build stuff17:05
cyphermoxanyway, AFAIK17:05
didrockscyphermox: I think cjohnston did the upstream merger part17:05
cyphermoxthen the proposal should be able to merge fine17:05
cyphermoxI'm not denying that the second part is necessary, just that it's not what's blocking the merge request17:06
cyphermoxdidrocks: do you happen to have magic ubuntuone access to be able to update the packaging branch for ubuntuone-credentials?17:06
dobeycyphermox: what are you suggesting?17:06
cjohnstonI did deploy-cupstream2distro-config17:07
cyphermoxcjohnston: ok17:07
cyphermoxin this case, the merge should be good to land17:07
didrockscyphermox: ah, I don't17:07
cyphermoxthen we still need to do cu2d-update-stack for the actual daily build17:07
cyphermoxdidrocks: actually17:08
cyphermoxI think I might have the right access, going to try17:08
dobeycyphermox: what are you "going to try" exactly?17:08
dobeyplease don't touch any ubuntuone branches17:08
cyphermoxupdating the config17:09
cyphermoxah, that oneś daily-release false too17:10
cyphermoxIt's already good17:10
dobeyyes, the change was just to remove the packaging branch, so ps jenkins would pass the tests again17:11
cyphermoxcjohnston: you should jsut be able to kick the merge proposal to try the build again, but I don't know how you do that17:11
dobeyhttp://s-jenkins.ubuntu-ci:8080/job/ubuntuone-credentials-ci/69/rebuild17:12
dobeysomeone just needs to rebuild it, if the live config is updated now17:12
cjohnstonrebuilding17:12
dobeylooks like it failed again with the same issue.17:15
dobeyso i guess the liv config isn't updated17:15
dobeylive even17:15
=== Ursinha-afk is now known as Ursinha
cyphermoxcjohnston: you'll need to check the config in http://s-jenkins.ubuntu-ci:8080/job/ubuntuone-credentials-trusty-amd64-manual/14/console to see why it still tries to grab debian, I'm not sure if that's covered by packaging_branch but regardless, it's not a process I know about, it's not cu2d.17:20
cjohnstonbzr branch $packaging_branch debian17:21
dobeymaybe jenkins needs a restart to refresh the config or something?17:22
dobeynot sure how that works17:23
cjohnstonthat would be very unfortunate17:23
cjohnstondobey: hrm17:23
cjohnstonif I just did a rebuild, then that would still have the packaging branch17:23
dobeydid the job config not change?17:24
cjohnstonit has a packaging branch field, but it is blank17:25
cjohnstonbut on a rebuild, it takes the config of the build you are rebuilding17:26
cjohnstonso I think it needs a new build17:26
dobeyhmm, maybe. i guess trigger a new build of the job?17:26
cjohnstonor the packaging branch to be removed when doing a rebuild17:26
cjohnstonit still requires a packaging branch17:27
dobeyi thought "rebuild" just triggered a new build of the same job17:27
cjohnstonit does, but it carries all the env vars17:28
cjohnstonhttps://code.launchpad.net/~dobey/cupstream2distro-config/u1creds-inline-packaging/+merge/197230 doesn't seem like the correct fix for the problem dobey17:28
cjohnstonit still expects and external branch for the debian files17:28
cyphermoxit's probably jsut carrying things over17:30
dobeywhy is it expecting one?17:30
cyphermoxwhat if you resubmit the merge proposal?17:30
cjohnstoncyphermox: dobey that wont work...17:31
cjohnstonall that the MP did was take away the default packaging branch17:31
cjohnstonit didn't remove the need for having a packaging branch17:31
dobeyand what specifies that need?17:32
cjohnstonyour funny17:32
cyphermoxcjohnston:        ubuntuone-credentials-trusty-amd64-manual: template: False ??17:32
* cjohnston has no idea.. still looking17:32
cyphermox(just guessing)17:32
cyphermoxI know fginther had a google doc about these things17:33
cyphermoxcjohnston: do you need a link to the document?17:37
cjohnstoncyphermox: the config in jenkins has:17:37
cjohnston# This project uses a packaging dir that is not compatible with pbuilderjenkins17:37
cjohnstonbzr branch $packaging_branch debian17:37
cjohnstonbut I don't know where that comes from17:37
cjohnstonyes please17:37
cyphermoxright, that's because of the separate debian/ branch17:37
cyphermoxcjohnston: from what I can see, you'd just need to remove that job in configurations and possibly also remove the part about trusty-amd64: False to have things use the standard process17:38
cyphermoxhttps://docs.google.com/a/canonical.com/document/d/1vZf1jxnP0XYrVSoJF3UoTPNhZBYtX9oQufpR8bdaDzQ/edit#heading=h.c8d6w5jt4m6017:38
Laneycjohnston: Could you retry http://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-glib2.0/22/ one more time? There was some kind of bug in the job that made it pull the wrong version of the tests (pinged jibel about it already)17:38
cyphermoxcjohnston: ^ it's not a very clear document though17:38
* cyphermox goes to get lunch17:38
cjohnstonLaney: done17:40
Laneythanks17:40
Laneysilly infrastructure :-)17:41
dobeyhmm, there is no indication in that document of what "template: False" actually means there17:42
cjohnstonim wondering if you do template false if you are using a seperate packaging branch17:47
cjohnstondobey: cyphermox I'm at a complete loss17:50
cjohnstonI'd say at this point we will need to wait for Monday17:50
cyphermoxwell, tbh I think it's just avoiding to use a template at all, and instead uses the job named whatever the title is there17:50
cyphermoxin this case, ubuntuone-credentials-trusty-amd64-manual17:51
cyphermoxcjohnston: you could just as well remove the config to try and put it back17:51
cyphermoxif things don't work as you expected17:51
dobeycyphermox: but it's not clear to me if that should be removed exactly17:51
cyphermoxdobey: most jobs don't need any of this configuration, and don't have a configuration: stanza at all17:52
cyphermoxthis was only required because you used merge mode for the bzr branches17:52
dobeycyphermox: right. and ubuntuone-credentials will still be merged by the u1 tarmac17:52
dobeycyphermox: which is why it's not clear if just removing that job is the right thing to do17:52
cyphermoxtarmac is irrelevant there17:52
cyphermoxanyway, the worst that can happen is that you'd have another message from jenkins on the merge proposal that says the attempt failed17:53
dobeycyphermox: only under the assumption that this config isn't what's preventing the CI autoland job from merging approved branches in that project17:53
* cjohnston lunches17:53
dobeycyphermox: and i don't see any clear indication if that is true or not17:53
dobeyanyway i also need to get lunch17:53
dobeybbiab17:54
cyphermoxany branch approved that hasn't been merged yet would still need to be reviewed by jenkins first anyway17:54
cjohnstonLaney: all better?18:10
Laneycjohnston: seems decent now, cheers18:13
cjohnston:-)18:13
psivaaseb128: btw with libglib2.0 (2.39.1-0ubuntu2) from -proposed the NetworkManager crash is not occurring anymore.18:22
seb128psivaa, I know, but thanks for confirming ;-)18:23
seb128ogra_, psivaa, asac: the fixed glib is in trusty proper btw18:35
seb128it just migrated there18:35
dobeyback19:11
dobeycyphermox: right. what i want to avoid though, is two separate jenkins trying to merge/commit the same branch into trunk19:14
dobeyi wonder if just changing this to use "autolanding: False" instead, would solve it19:39
cyphermoxnope, that's something else19:52
cyphermoxthat will make the merge proposals not merged at all19:52
dobeyhow so?20:05
dobeyand does not merged mean no ps jenkins running tests?20:05
dobeybah. i guess i'll just go all in, instead of trying to stage this migration in a reasonable way20:21
asac19:35 < seb128> ogra_, psivaa, asac: the fixed glib is in trusty proper btw20:36
ogra_ok20:44
ogra_=== Image r39 building ===20:44
popey\o/20:44
asacoh cool20:55
dobeycyphermox, cjohnston: https://code.launchpad.net/~dobey/cupstream2distro-config/u1creds-fix/+merge/197264 <- maybe we could get this merged and deployed now then? i've pulled the landing of that project's trunk out of the u1 tarmac setup now.21:11
dobeycihelp ^^21:26
dobeywell, time for me to head off. guess this won't get merged/deployed today. :-/21:58
dobeylater21:58
ogra_=== Image r39 DONE ===22:03
ogra_(since a while)22:03
asacoh :)22:03
asacogra_: thanks22:03
ogra_np :)22:04
* ogra_ goes afk again ... 22:05
* popey flashes22:13
asacpopey: is it any good?22:28
asacdashboard is still greenish :) at 11 pass22:28

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