/srv/irclogs.ubuntu.com/2013/11/04/#ubuntu-ci-eng.txt

sil2100jibel, vila, Mirv, asac: hi!08:24
sil2100jibel, vila, Mirv, asac: I see that the otto cu2d check jobs are failing again with the Could not resolve 'archive.ubuntu.com' error08:24
sil2100Do anyone of you know what's up with that again? New DNS issues?08:24
sil2100retoaded: ^08:24
vilasil2100: url ?08:25
sil2100For instance, http://10.97.0.1:8080/job/autopilot-trusty-daily_release/396/label=qa-intel-4000/console08:25
sil2100On both nvidia and intel08:25
Mirvsil2100: hi!08:26
sil2100(so http://10.97.0.1:8080/job/autopilot-trusty-daily_release/396/label=autopilot-nvidia/console as well)08:26
vila/var/log/syslog: Nov  4 00:08:25 trusty-i386-20131104-0008 kernel: [86262.203773] type=1400 audit(1383523705.376:356): apparmor="DENIED" operation="open" info="Failed name lookup - disconnected path" error=-13 parent=26433 profile="/sbin/dhclient" name="etc/ld.so.cache" pid=26437 comm="dhclient" requested_mask="r" denied_mask="r" fsuid=0 ouid=008:26
Mirvsil2100: I don't know what's up this time, but the archive.ubuntu.com is not (again) reachable from magners either so maybe some firewall reconfiguration again?08:26
sil2100Mirv: yep, apparmour issues again...08:27
sil2100vila: what did we do last time to fix this actually...? ;)08:27
vilaMirv: wget archive.ubuntu.com works08:27
vilaMirv: works from magners, that's the only reliable check I know of, forget the others08:28
sil2100As we had this before, exactly the same issue I gues08:28
Mirvvila: ah, you're right, the firewall apparently just blocks ping08:28
vilaMirv: yes, for unknown IS reasons (probably good ;)08:28
Mirvvila: ping works from my own machine so I didn't think it'd be different. but right, so apparmor agian.08:28
vilaMirv: yes, I learned it the hard way too08:28
sil2100Everything is crimson red because of this sadly08:28
vilasil2100: probably not exactly the same, related to apparmor again though08:29
vilasil2100: same root cause as far as CI is concerned: a user job breaks the infrastructure because otto *requires* host and lxc to be up to date08:29
sil2100vila: *sighs*08:32
vilasil2100: yeah, don't tell me ;-/08:33
vilasil2100: not to mention all upstream-merger are stucked :-/08:33
vilasil2100: so, back to what was done last time: ping stgraber, he's the most efficient to diagnose the lxc/apparmor issues lately08:34
veebersvila: another question for you. I'm getting this error on the generic-mediumtest-runner-mako: qmlscene: could not find a Qt installation of ''08:35
vilasil2100: may be check if some updates occurred on apparmor or lxc ?08:35
vilaveebers: you want to talk to autopilot devs to provide us a better error message ;)08:35
vilaveebers: that's not the first time I see this undecipherable one08:36
veebersvila: heh, what I'm wondering is why /usr/bin/qmlscene doesn't work on that device?08:36
veebersthat is the direct error message from running /usr/bin/qmlscene -testability /path/to/qml/file08:36
vilaveebers: I didn't track what caused it the previous times sorry :-(08:36
veebersit works find on my device after a phablet-flash08:36
sil2100vila: will do ;)08:37
vilameh08:37
vilasil2100: not to mention all upstream-merger *maguros* are stucked :-/08:37
jibelvila, sil2100 from networking.log dhclient: error while loading shared libraries: libc.so.6: cannot open shared object file: Permission denied08:54
vilajibel: even worse ;)08:54
jibelvila, can you downgrade the kernel and see it makes the error go away?08:56
jibelon the host08:56
vilajibel: not yet, I'm restoring access to upstream merge maguros08:59
jibelvila, sil2100 this problem start on Nov. 1rst with kernel 3.12.0-1, rebooting with 3.11.0-12.19 should work. If it does ping the kernel team09:03
didrocksgood morning09:05
didrockssil2100: Mirv: psivaa: popey: meeting time!09:31
psivaadidrocks: joining09:31
didrockssil2100: just waiting for you now ;)09:32
didrockssil2100: not around? you shouldn't suffer from jet lag09:40
sil2100Crap!09:42
sil2100I got used to NOT having the meetings09:42
Mirvdidrocks: the build probs should be have a mention at https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AuDk72Lpx8U5dHFtUmlPOUtCRk8zR2dtaEpIbUVhMmc#gid=309:43
sil2100jibel, vila: we had the same issue last time with dhclient09:44
sil2100jibel, vila: I remember dhclient had problems also loading libc09:44
jibelsil2100, right but can you try to boot kernel 3.11, the incident began with 3.12 then notify the kernel team?09:45
Mirvsil2100: vila: I checked the other uploads and there are no dbus/lxc/apparmor uploads, so as jibel says downgrading the kernel would seem the first thing to try09:46
ogra_=== Image #10 promoted ===09:51
didrocks\o/09:52
ogra_popey, promoted  10/20131031.109:52
ogra_=== Image #11 building ===09:52
didrocksgreat!09:53
ogra_hmpf ... compiz doesnt get the layer ordering right anymore ... funny ... all menus, tooltips and popups are behind all other windows ... even pulldowns in the browser09:54
sil2100\o/09:58
popeyogra_: super10:00
didrocksvila: hey! I tried to get in touch with doanac and it seems you didn't tell him that the execution test service should use the same parameters than the otto jobs?10:02
didrocksvila: do you take care of explaining what every parameters do or should I?10:02
viladidrocks: err, surprising...10:20
didrocksvila: you meant, you talk about them with him?10:21
viladidrocks: unless you mean the parameter *names* were different, I'm pretty sure the discussions mentioned them all10:21
didrocksvila: names and syntax are different10:21
didrockslike the tests is not the command running autopilot tests10:21
vilaright, but the features are there right ? So it should be a matter of adapting both sides to agree no ?10:22
didrocksvila: well, that's what I discussed with you. That we can quickly plug that in and so adapt the parameters to what we have for now10:23
didrocksthe semantic for some parameters are unfortunately different, I see a bigger road block on that one10:23
viladidrocks: right, so, since I don't know enough of the specifics on both sides to act as a proxy, better get in touch directly. In the midlle/long term though, I think we want to better separate setting up the environment and running the tests, it seems weird to have to implement the later differently in different tools10:25
didrocksvila: interesting, not sure why we had that meeting where I needed to explain to you all parameters though. Will talk with him directly then10:25
viladidrocks: so that I start to understand how cu2d works, I didn't have the same one with doanac and I thought the parameters and their intent were clear enough for him to ensure it will be providing the same semantic10:26
didrocksno worry, I'll fix that with him directly. I may have misunderstood the goal of this meeting apparently10:27
viladidrocks: I reported that meeting to the whole team and we even had a session were we discussed it...10:27
viladidrocks: correction, wasn't the whole team but only doanac and fginther10:35
didrocksok10:35
viladidrocks: notes here https://docs.google.com/a/canonical.com/document/d/1ISclO4aN2CQ2vJ5PeryF2gt81lqX-5_-mmGADNrK04o/edit#heading=h.3nv3w86usggi but further work hasn't been tracked there :-/10:38
viladidrocks: so the overall feedback *I* got was that the features were ready or almost ready and my understanding was that some tweaks were required to have matching parameters. But I wasn't involved in the testing/coding so I din't (still don't) know the exact state.10:41
didrocksvila: I just opened the proposed job that was told ready and didn't see any evolution on the parameters10:42
didrockscompared to when we first discussed about it10:42
vila:-/10:42
ogra_=== Image #11 done ===10:56
didrocksgreatness!10:59
didrockssil2100: http://bazaar.launchpad.net/~ubuntuone-hackers/ubuntu-download-manager/trunk/revision/151 can you SRU just that commit please?10:59
didrocksto saucy10:59
didrocksthis is to avoid getting people on saucy in trouble10:59
didrocks(for the next 2 weeks)10:59
didrockssil2100: old landing ask 266 (it's the short term fix)10:59
sil2100didrocks: ACK11:04
sil2100didrocks: ok, I see it's in saucy's trunk, so I guess I can use cu2d to get it in, right? There is only this commit in saucy trunk waiting for release11:07
popeyogra_: #11 I can't make phone calls. I can receive though11:07
ogra_popey, ouch11:07
popeyhmm, cant send texts either, wonder if I have run out of credit11:08
sil2100didrocks: anyway, taking care of that this instance11:08
didrockssil2100: just do what you prefer :)11:11
vilasil2100: qa-intell-4000 back to previous kernel, seems like otto is happy (minus the fact that it's supposed to always run in sync with the latest kernel....)11:14
sil2100vila: \o/11:14
vilasil2100: I tweak /etc/default/grub and that will need to be reverted when you have a fix11:14
vila*tweaked11:15
sil2100vila: well, at least we have tests running, too bad that with not the latest kernel but we'll switch to that once it's fixed properly11:15
vilasil2100: or otto_setup is run and breaks again... I think my tweak is not robust enough against that11:15
popeyogra_: phew, credit ran out11:27
ogra_phew11:28
popeyogra_: #11 seems good on mako11:33
ogra_great, still upgrading here11:34
vilasil2100: dx-autopilot-nvidia also reverted (with a better tweak also applied to qa-intel-4000)11:37
sil2100vila: will it still get broken again when otto_setup is ran?11:38
vilasil2100: a job spontaneously triggered on qa-intel-4000 which served as a test, can you trigger one for nvidia ?11:38
vilasil2100: it should stay on the same kernel but hey, we're in production, I couldn't test as much as I'd like there11:39
sil2100Let me trigger a quick one11:39
vilasil2100: so we'll need to revert the line added in /etc/default/grub, I've put a comment there so that should be easy11:39
* vila lunches &11:40
sil2100Thanks! :)11:40
sil2100Running the test run in the meantime11:40
ogra_popey, maguro seems fine on image 1112:08
ogra_(well SMS and calls do)12:08
sil2100vila: all works ok it seems \o/12:29
ogra_BAH !12:30
ogra_all unity8 tests fail again on mako12:30
ogra_this is getting tiring :(12:30
sil2100Mirv: https://code.launchpad.net/~sil2100/cupstream2distro-config/click_stack_for_saucy/+merge/193753 <- can you take a look? We created a saucy branch for this one12:31
psivaaogra_: i think plars pushed a fix for turning the screen on for unity8, still failing, the reason appears different. could be temp, i am rerunning12:32
ogra_psivaa, yeah, i remember the fix worked in image #1012:33
psivaaogra_: yes, the error message is different this time compared to the failures pre image 1012:34
psivaaohh no some of them are the same :(12:35
ogra_the pattern looks the same at least12:35
psivaayea12:36
=== Ursinha_ is now known as Ursinha
psivaathere is a powerd crash in the runs with image 11 that i dont remember seeing with 1012:40
psivaaand the fix is using powrd-cli so that may explain why that dint work12:41
ogra_yeah12:42
ogra_do you run powerd-cli as root ?12:42
ogra_iirc you need to12:42
Mirvsil2100: ok12:46
psivaaogra_: powed-cli is running in the device now, but not sure if that was running whilst unity8 tests were running. will watch that12:50
sil2100hmmmm12:51
=== alan_g is now known as alan_g|lunch
sil2100Mirv: good catch! Modified13:04
=== josepht changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: josepht | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
Mirvsil2100: top-approved now.13:23
sil2100Mirv: thanks!13:36
plarspsivaa: odd, I got it to pass just fine on Friday13:49
ogra_plars, yeah, i wonder if the code was reverted or somethign13:49
psivaaplars: i think the issue is powerd crash this time but still yet to rerun13:50
plarsI'll take a look13:50
psivaawaiting for all the tests to complete13:50
psivaaplars: ack13:50
plarsugh, more rain13:50
=== alan_g|lunch is now known as alan_g
sil2100didrocks: do you know if we're dropping powerpc support in overall?14:07
sil2100Like, completely?14:07
fginthermorning14:08
sil2100Morning14:08
cjwatsonsil2100: No14:09
sil2100cjwatson: ok14:09
sil2100Then I guess I'll have to revert robru's change in libunity-webapps14:09
cjwatsonUh14:09
cjwatsonYou said "completely" which is a very broad question14:10
cjwatsonIt may be correct to drop it in specific cases where it wouldn't be buildable anyway14:10
sil2100cjwatson: well, here it would mean that suddenly unity7, bamf, and many many unity components would have to disappear for powerpc ;)14:10
cjwatsonI don't know about the case of libunity-webapps14:11
cjwatsonIt's generally best to keep things Architecture: any unless there's a strong reason otherwise14:11
cjwatsonDo you mean the webbrowser-app dependency?14:12
cjwatsonYou could always make that architecture-specific14:13
cjwatsoni.e.  Depends: webbrowser-app [amd64 armhf i386]  Suggests: webbrowser-app [!amd64 !armhf !i386]14:13
cjwatsonOr actually just drop the Suggests part of that, I guess, since it's unsatisfiable anyway14:14
sil2100Actually, we were already Suggests webbrowser-app to workaround this14:14
cjwatsonYeah, but now you want to hard-depend on it on architectures where it's available14:14
cjwatsonJust Depends: webbrowser-app [amd64 armhf i386] would be fine for that14:15
sil2100Right, or webbrowser-app [!powerpc] I guess?14:15
cjwatsonNo14:16
cjwatsonWe have more than four architectures14:16
cjwatsonIt's more correct to specifically enable it on the architectures where we know webbrowser-app is available14:16
sil2100Yes, but we only ban it from this one particular?14:16
cjwatsonBut that would be wrong!14:16
sil2100Ok ok14:16
cjwatsonwebbrowser-app | 0.22+13.10.20131011.1-0ubuntu1 |        trusty | source, amd64, armhf, i38614:16
cjwatsonWhich does not include arm64 ...14:16
cjwatsonNor the further architecture we'll be adding in the next couple of months14:17
sil2100Right, but I guess it will be supported as soon as it's possible, so then we'll have to update the package again, right?14:17
cjwatsonarm64 doesn't work right now for the same reason powerpc doesn't work14:17
cjwatsoni.e. v8 not ported14:17
cjwatsonNo particular reason to expect either one to work before the other AFAIK14:18
sil2100Ok, I'm fine with any solution, just wanted to find the one that would be less troublesome to maintain :)14:18
sil2100But I guess this will be the correct one here14:18
cjwatsonIt makes more sense from an archive point of view to enable the dependency only where it's specifically available14:18
cjwatsonHopefully the v4 interpreter in Qt 5.2 will be more portable14:18
sil2100Thanks for the pointers!14:21
cjwatson(It's supposed to be, from what I hear)14:21
plarsogra_, psivaa: given how good the custom image looked at http://10.97.0.1:8080/view/Touch/view/Ubuntu%20Touch%20Master%20Jobs/job/trusty-touch_custom-mako-smoke-master/ I'm surprised unity8 and webbrowser had problems on the mako image. I'm rerunning now, expect they will pass14:22
plarsunity8 is looking fine for me locally so far14:22
ogra_well, lets see how the next image behaves ... probably a temporary glitch14:23
plarsogra_: is 12 building now?14:24
plarsogra_: I just restarted some tests on 1114:25
ogra_plars, nope14:25
ogra_earliest after the next meeting14:25
plarsogra_: ok, I'll keep pushing on 11 then :)14:25
plarsogra_, popey: unfortunately we started having to handle a LOT of extra device-not-found errors, which mostly seem to be attributed to the mtp/adb conflict14:39
ogra_plars, the lxc-android-config package with the new upstart job from sergiusens is the first step towards fixing this14:40
sil2100cjwatson: could you maybe review this branch (if you have a minute and the permissions), since the upstream dev is not around right now: https://code.launchpad.net/~sil2100/libunity-webapps/webbrowser-app_depends_arch/+merge/19378514:40
ogra_(uploaded today)14:40
plarspopey: but if you look at the results of filemanger, it seems to always be up and down between 3 or 4 failures and 19 or 20 failures14:40
plarsogra_: that's good to hear :)14:40
cjwatsonsil2100: I added an Approve review, but it shows as "(community)" so it probably doesn't count for landing14:44
cjwatsonIt can serve as your core-dev +1 though :)14:44
sil2100;)14:46
sil2100Thanks!14:46
doanacvila, didrocks: let me know how we should resolve the parameters for this thing. I think we are already close to a 1-to-1 mapping of the parameters, but i need to know if theres something big missing15:19
didrocksdoanac: sure, want to chat about it?15:22
doanacdidrocks: sounds good. how about in 8 minutes? (half past the hour)15:22
didrocksdoanac: good for me as well :)15:22
vilaMirv: <mlankhorst> vila: enjoy, I found a fix for glamor-egl :P15:28
vila<mlankhorst> https://bugs.freedesktop.org/attachment.cgi?id=8861915:28
doanacdidrocks, vila: i'll create a hangout for us15:29
doanacvila, didrocks: http://10.97.9.20:8080/job/trusty-touch-mako/17/15:36
doanacvila, didrocks: http://10.97.0.1:8080/job/autopilot-trusty-daily_release/302/parameters/15:37
didrocksdoanac: vila: http://10.97.0.1:8080/job/autopilot-trusty-daily_release/201/label=autopilot-nvidia/console15:43
balloonsping fginther15:56
=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
tedgIt seems that dbus-test-runner is configured to do CI on Quantal as well as Trusty.16:25
tedgDoes anyone know why that could be?16:25
tedghttp://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/view/head:/stacks/head/qa.cfg#L9816:25
=== josepht changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
plarsdidrocks: no landing meeting?17:05
tedgev, Where should I file the bug that errors.u.c misspells "optimized" in "optimized out" ?17:05
didrocksplars: oh sorry, we told no meeting today, will resume tomorrow afternoon17:05
* ogra_ will be late, running meeting is running over 17:05
plarsdidrocks: I missed that I guess, sorry17:05
ogra_oh, cool17:05
didrocksplars: I missed to ping you, my fault, sorry ;)17:05
didrocksI pinged the others :)17:05
=== bfiller is now known as bfiller_afk
psivaaplars: are you doing anything with the smoke maguro in the lab, can i rerun filemanager and gallery?17:40
psivaaballoons: we are still seeing filemanager test failures due to too many entries in /home/phablet/ .17:41
plarspsivaa: go for it, I've mostly been focusing on mako at the moment17:41
psivaaplars: ack17:42
balloonspsivaa, we're still trying to land a couple mp's to work on this..17:42
balloonspsivaa, https://code.launchpad.net/~vthompson/ubuntu-filemanager-app/fix-carlos-model/+merge/19378317:42
psivaaballoons: ack, thanks :)17:42
balloonspsivaa, the paste bug is fixed by that.. then I can land my big branch17:42
balloonsand then hopefully it's all fixed :-)17:43
=== alan_g is now known as alan_g|EOD
cjwatsonFYI: publisher down pending investigation of I/O errors18:01
=== Ursinha-afk is now known as Ursinha
=== fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: fginther | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
fgintherjosepht, any vanguard work I should know about?18:39
josephtfginther: not a thing18:43
fgintherballoons, pong18:50
fgintheralesage, do you know why dbus-test-runner is still configured to build on quantal?18:52
alesagefginther, no sir18:52
alesagefginther, propose to amend18:52
fgintheralesage, ack, I'll drop it, tedg is hitting build failures to do it18:52
alesagefginther, thank you18:52
balloonsfginther, https://code.launchpad.net/~vthompson/ubuntu-filemanager-app/fix-carlos-model/+merge/193783. qt5-proper ppa doesn't have a trusty build, but the virtual test enviroment is using it18:54
fgintherballoons, looking18:56
tedgfginther, thanks!18:57
=== bfiller_afk is now known as bfiller
fginthertedg, your dbus-test-runner MP is passing now19:33
fgintherballoons, https://code.launchpad.net/~vthompson/ubuntu-filemanager-app/fix-carlos-model/+merge/193783 is passing now19:34
balloonsfginther, ty sir :-)19:35
tedgGreat!  Thanks fginther!19:36
veebersfginther: hey how are you, query: Would you know why `/usr/bin/qmlscene` works on my local device but doesn't on the jenkins devices? i.e. I'm getting this error: qmlscene: could not find a Qt installation of ''20:12
fgintherveebers, hi, it could be missing a qt plugin dependency20:13
fgintherveebers, what job is this?20:13
veebersfginther: this is an example: http://10.97.0.26:8080/job/generic-mediumtests-runner-mako/2987/artifact/results/autopilot/sudoku_app_results.xml20:16
fgintherveebers, when you run them on your device are you using the deb packages to install and test?20:28
veebersfginther: In my test last night yes. But I've just re-flash my device so I'm going to try something20:37
=== Ursinha-afk is now known as Ursinha
veebersfginther: on my locally freshly flashed device when running `/usr/bin/qmlscene` I don't get that error20:45
fgintherplars, is there a reason sudoku_app tests are not run for the smoke testing?20:51
plarsfginther: I think just nobody asked for them20:52
plarsfginther: Are there tests that work?20:52
fgintherplars, so far they don't20:53
plarsfginther: then that's probably why nobody asked for them :)20:53
plarspopey: can't seem to get filemanager down under 10 failures on mako, but on maguro it only had 3 today22:51
plarsalso weather app is considerably worse under mako for some reason22:52
xnoxis there a bug tracker for CI infrastructure issues?23:02
xnox(bug reports / wishlist items / etc?)23:02
robrufginther, can you check why jenkins didn't run here? https://code.launchpad.net/~justinmcp/unity-chromium-extension/mediaplayer-init-fix/+merge/193549 guy's a new employee, might need to be added to some kind of whitelist?23:02
fgintherrobru, yep, that's probably the case, I'll get the list updated23:04
robrufginther, thanks23:04
robrufginther, what's the recommended way of handling community-submitted branches? obviously jenkins can't run on them by default, but is there some one-off way of saying 'jenkins should run for this one branch'?23:09
robruor do i have to steal the branch and take all the credit with my own MP?23:10
fgintherrobru, the only way to do this right now is to manually trigger the job from the jenkins UI23:10
robrufginther, hrm, can you link me to that part of the jenkins maze?23:10
fgintherrobru, you need to find the ${lp_project}-ci job here: http://10.97.0.26:8080/23:16
fgintherrobru, for example: http://10.97.0.26:8080/job/unity-chromium-extension-ci/23:16
robrufginther, ahhhh, that's a different jenkins! I don't have a login to be able to kick jobs there23:17
fgintherrobru, if you create one, I'll add build permission23:18
robrufginther, oh, hrm, it says 'robru' is taken, but doesn't have any kind of password recovery button...23:19
fgintherrobru, I can reset it23:19
robrufginther, ok, thanks. for the cu2d jenkins that one just lets me sign in with SSO...23:20
fgintherrobru, you should be setup now23:21
robrufginther, great, thanks!23:22
fgintherrobru, when building the -ci job, you'll need to supply the landing_candidate (the lp:blah branch name) and the merge_proposal (the URL of the MP in launchpad)23:22
=== fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
=== Ursinha is now known as Ursinha-afk

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