/srv/irclogs.ubuntu.com/2014/06/03/#ubuntu-ci-eng.txt

kgunnanyone around to provide some silo love ?...so we can start building for mir.01:35
kgunnmaybe Mirv when you come on? ^01:36
imgbot=== trainguard: IMAGE 62 building (started: 20140603 02:10) ===02:10
imgbot=== trainguard: IMAGE 62 DONE (finished: 20140603 03:35) ===03:35
imgbot=== changelog: http://people.canonical.com/~ogra/touch-image-stats/62.changes ===03:36
Mirvkgunn: assigned and kicked a build: https://ci-train.ubuntu.com/job/landing-008-1-build/83/console04:24
plarsogra_: 61 and 62 still have dbus-x11 installed it seems, so they are still going to be broken. I guess from the suggests in dbus? That's the only rdepends I can find for it that we have installed04:41
Mirvplars: I can see there's a new line for reverting the split greeter, I'm not sure if that's because this previous fix failed to work or if it's just in case it's needed04:49
Mirvplars: Suggests: does not bring a package in, Recommends would04:49
Mirvso, weird04:50
Mirvoh, right, so it's in case the dbus split wouldn't be done04:56
Mirvit does not seem that at least simply again installing dbus-x11 and removing /etc/X11/Xsession.d/75dbus_dbus-launch would make #62 show more than a black screen for me06:02
Mirvit might be of course also because of my 5.3 experiments so I could need a fresh wipe06:04
MirvI'll assign a silo and kick a build of the revert anyhow, better to have it ready06:21
ToyKeeperNice.  Image 62 crashed four times after flashing, before I even turned the screen on to try it...  and all the indicators are missing.06:36
Mirvinterestingly, now that I upgraded my #62 + dbus-x11 - 75dbus_dbus-launch to the Qt 5.3 PPA, everything seems working correctly including indicators06:39
Mirvso maybe it was that my black screen was because of non-clean wipe (I've seen it before), but that restoring dbus-launch binary makes things work. or alternatively Qt 5.3 is simply that much better :)06:40
Mirvnote that the unity8 rebuild does have the split greeter etc anyhow, everything's up-to-date.06:40
ToyKeeperThis was a fresh flash.  Maybe it was a boot fluke, but it's not a good sign when the phone tries to melt itself on first boot.06:45
ToyKeeperIt was getting hot enough that I worry it could cause damage...  so I shut it off completely for now.06:46
ToyKeeperEr, and one of the crashed processes was /sbin/init ... which really worries me.06:55
MirvI seem to have /sbin/init .crash file from last Wednesday, so nothing totally new apparently06:58
* sil2100 sighs07:10
pete-woodscihelp: can anyone help me figure out what happened to the packages in silo #009? https://ci-train.ubuntu.com/job/check-publication-migration/17381/console07:17
pete-woods"WARNING hud (14.04+14.04.20140528-0ubuntu1) is in no known space (and time). We didn't find it anywhere. It could have been rejected, or you run it too quickly and it's not published in proposed yet, or a network issue happened. If this persistent some time after the publication, you need to ping a landing team member to get more informations."07:17
sil2100ogra_: ok, I understand from the e-mails that the current unity8 and such are causing indicators not appearing, but I wonder why some application tests still fail completely07:17
sil2100pete-woods: I will try looking at this, wanted to do it yesterday already but there was too much chaos07:18
sil2100pete-woods: since the package is in -proposed now when ai last checked07:18
sil2100pete-woods: actually, wait, no07:19
pete-woodssil2100: that one was a failed SRU07:19
sil2100pete-woods: ok, this needs some intervention07:19
pete-woodssil2100: just let me know if I need to change anything07:20
Mirvsil2100: read the backlog from #ubuntu-touch if you didn't yet, regarding pete's landing07:22
Mirvit was probably filtered out ie never copied to archives07:22
sil2100Mirv: will do that and poke some archive admins for help here07:22
Mirvsil2100: poking was also done, so if you combine the #ci-eng and #touch discussion you should have the whole picture I think07:23
Mirvso probably you'll need to let didrocks know if you want that filter changed or something else, and then probably republish is needed07:24
Mirvsil2100: http://irclogs.ubuntu.com/2014/06/02/%23ubuntu-ci-eng.html#t10:42 + aha we don't have #ubuntu-touch logs??07:25
Mirv+ http://pastebin.ubuntu.com/7578497/ then07:26
sil2100Mirv: thanks! I have this in my logs at least... but hm, why did it suddenly use the sru-staging PPA?07:28
sil2100I didn't configure it to do so07:29
sil2100I'll check the code and see what's going on07:29
sil2100Mirv: so, I'll have to chat with didrocks again, since it seems to be the default behavior so it should have done the same the first time the SRU was published07:32
sil2100Mirv: we don't have logs of that though since it was long ago07:33
sil2100Mirv: so not sure why suddenly now it would get 'filtered out'07:34
Mirvsil2100: ok, sounds good (to have a chat). I thought the abandoning of sru-staging was something you had agreed to last week or such.07:34
Mirvso I didn't have much idea what should be the case07:35
sil2100It was nothing that has been discussed with me around at least07:35
=== iahmad is now known as iahmad|afk
ogra_sil2100, ugh ... that looked totally different last night, 61 was mostly green07:51
sil2100ogra_: looking at 61 now it's similar though, makes me feel a bit uneasy07:53
ogra_sil2100, yeah, dropping dbus-x11 didnt help at all it seems07:54
ogra_sil2100, see the mail discussions though07:54
ogra_sil2100, it seems we need 50% of dbus-x1107:55
sil2100Yeah, saw that discussion more or less07:55
ogra_(dbus-launch, but not the Xsession.d part)07:55
ogra_and pitti seems to have a good suggestion for a fix by dropping a config option07:56
sil2100ogra_: right, I see that now - need to read up about that option though07:59
sil2100ogra_: ah, so this way the dbus session won't launch even with dbus-x11 installed?07:59
Saviqo/08:00
Saviqsil2100, can you kick https://launchpad.net/~ci-train-ppa-service/+archive/landing-019/+build/6064416 ? I hope for a flaky test - only failed on amd6408:00
ogra_sil2100, gdbus uses dbus-launch to fire up a dbus when it first tries to connect and finds one08:01
ogra_sil2100, the first gdbus call seems to happen before the Xsession script is processed08:02
ogra_this then starts another dbus-daemon08:02
sil2100Saviq: sure08:06
ogra_hmm, no answer from bfiller at all ... so we have to roll back i guess08:08
ogra_:(08:08
sil2100He'll be like in around 5 hours08:09
sil2100Which is quite long...08:09
sil2100It's really a shame though, such a big landing reverted08:10
ogra_sil2100, well, pitti just proposed a fix for which we need to roll back the dep dropping and he needs to upload a fix to dbus ...08:10
sil2100ogra_: remember on Friday when I was saying about getting ready for this big thing landing and you were like "it's nothing risky, it was in testing for months!" ;) ?08:11
ogra_but we will have the same prob as yesterday08:11
MirvI've #62 working just fine with Qt 5.3, after installing dbus-x11 back in and removing the X related script :)08:11
ogra_needs to be tested in the lab08:11
ogra_Mirv, right08:11
ogra_Mirv, and thats what we tested with plars yesterday08:11
sil2100Right08:11
ogra_i didnt think of the fact that dbus-launch could play any role by being called directly08:12
ogra_which means we need exactly 50% of dbus-x1108:12
sil2100But what is the cause of all the current failures we have in smoketesting?08:12
ogra_but pitti agrees its a dbus bug that the Xsession part gets processed at all08:13
ogra_sil2100, that we completely dropped dbus-x1108:13
ogra_we need to keep half of it08:13
sil2100ogra_: wasn't is mentioned that 'smoketesting is fine after dropping dbus-x11' on the ML?08:13
ogra_or in other words, make the Xsession script a no-op if gdbus fired one up already08:13
sil2100Ok, so it's like, everything has one cause08:14
ogra_sil2100, yes, before i went to bed mako was all green (not all tests had run)08:14
ogra_http://paste.ubuntu.com/7578684/08:14
ogra_thats the proposed fix ... i will test that in a minute and pitti will upload it08:14
sil2100Let's wait for this to happen, the big revert silo might build itself in the meantime08:15
ogra_but we still have to make a decision if we want to risk it or complketely roll back08:15
sil2100So that we have a fallback in case of nothing helping08:15
ogra_the prob is that we tested with dbus-launch installed ... where it worked fine ... my change dropped that alongside08:15
* ogra_ gets his flo and fresh coffee ...08:15
sil2100I would try risking it, really, as such a big revert is never 100% sure as well08:16
ogra_then i'll test08:16
ogra_sil2100, i agree08:16
brendandcihelp - can someone help me understand these jenkins failures? https://jenkins.qa.ubuntu.com/job/generic-deb-autopilot-runner-mako/1027/console08:21
sil2100pete-woods: so, I need to change something in our CI Train code to get this moving08:23
sil2100pete-woods: will try to finish it up today, but there's a lot going on08:23
pete-woodssil2100: okay, thanks for the update :)08:23
vilabrendand: RO file system despite phablet-config writable-image, that's beyond my understanding, you want plars or doanac08:24
vilabrendand: is this a recent failure ?08:25
mandelsil2100, I forgot the channel of the train, can you telll it to me?08:26
sil2100mandel: #ubuntu-ci-choo-choo ? :)08:26
mandelsil2100, thx!08:26
mandelsil2100, malta has drained my brain08:27
brendandvila, semi recent. a specific mr keeps failing for that reason08:27
brendandvila, since the end of last week i think08:28
vilabrendand: add fginther to your ping then since it's a s-jenkins job08:29
vilabrendand: they are all in US08:29
brendandvila, i'll take one more try and see if it happens again08:29
popeysil2100: ogra_ landing meeting?08:30
sil2100popey: yep08:30
ogra_yeah, sorry, still in debug mode ... on my way08:31
* ogra_ cant connect ... 08:32
Saviqsil2100, ogra_, train is happy with revert silo, what do we do?08:34
ogra_Saviq, we have the landing meeting right now ... discussing that ... wanna join ?08:34
sil2100Saviq: come over, let's have a chat08:34
ogra_https://plus.google.com/hangouts/_/canonical.com/landing-meeting08:35
brendandcihelp - i have my vpn up but can't reach s-jenkins09:16
evbrendand: interesting. Having a look now09:17
sil2100brendand, ev: bfiller had a similar issue yesterday already09:17
sil2100It seems after relogging 10 times it started working...09:17
brendandev, some people mentioned similar yesterday09:17
evbeing able to reach it and being able to log in are different problems, no?09:18
vilabrendand: funny you mention that, I just had to fix my laptop and desktop configs following instructions from https://wiki.canonical.com/UbuntuEngineering/CI/VPN09:19
vilabrendand: I was using a config that didn't followed the updated instructions09:19
brendandvila, but i used this yesterday09:20
brendandvila, did it change in that time?09:20
vilabrendand: yeah, I used mine for months09:20
brendandvila, ok i'll check the config09:20
vilabrendand: I suspect that some pieces were left in place to let people update to the new (long term supported) bits so nobody realized they had to change the relevant bits09:21
vilabrendand: for NM based config I had to change to 10.99.244.1 for the ubuntu-ci DNS server (I was using an old IP)09:21
vilabrendand: for the openvpn based I had to stop using my hand-crafted scripts and switch to /etc/openvpn/update-resolv-conf for 'up' and 'down'09:22
brendandvila, so that seems to have done the trick09:22
brendandvila, but all i did was remove some whitespace for the up and down lines09:23
brendandvila, not even sure how the whitespace got there09:23
brendandbad copy and paste i guess09:23
vilabrendand: and you did 'service openvpn restart' ?09:24
brendandvila, no i just run openvpn from the command line09:24
vilabrendand: and you had a previous running but not giving you the right infos correct ?09:25
brendandvila, well yes09:25
vilabrendand: so may be it was relying on some old IP that got cleaned up recently09:25
brendandvila, perhaps09:26
vilabrendand: anyway, if it works now, I would keep that explanation in mind and ask retoaded for some confirming bits09:27
evthanks vila09:56
vilaev: you didn't have any issues with your vpn I suspect ?09:57
evnot at present10:18
vilaev: right, so you probably had an up-to-date config already10:19
sil2100pete-woods: ok, let's hope that now HUD goes forward10:33
* sil2100 did some hacks to push it without a rebuild10:33
sil2100pete-woods: yesss! hud re-appeared in the unapproved queue :)10:35
pete-woodssil2100: :D10:36
* Mirv gone for a while, will check irc etc still today11:49
sil2100Mirv: o/12:02
=== fginther` is now known as fginther
=== tedg is now known as ted
popeysil2100: was there not going to be another catch up?13:08
sil2100popey: yeah, let's wait for the image first though13:08
popeyk13:08
pmcgowanimgbot, stunt13:10
* imgbot rolls on its back and purrs13:10
ogra_:)13:10
popeyhah13:10
ogra_publisher has both packages in proposed now ...13:11
* pmcgowan waits for 6213:11
ogra_next publisher run is ours13:11
ogra_then we can start a build13:11
* sil2100 spams rmadison13:11
ogra_popey, i guess we wont have the image before the evening meeting ... at least not much ... and we all worked together over the last few hours anyway13:12
ogra_(in #ubuntu-devel mostly)13:12
ogra_so i dont think we need another meeting right now13:12
popeyok13:12
=== mhr3__ is now known as mhr3
* ogra_ is afk for a moment13:30
* Mirv back13:34
sil2100ogra_: it's in!13:40
sil2100ogra_: or wait13:40
sil2100hmmm13:40
sil2100ogra_: ok, so actually rmadison says it's still in -propsoed, but citrain thinks otherwise13:40
Mirvsil2100: citrain is too fast afaik13:43
sil2100I thought that with the latest changes it should work similar to how rmadison works13:43
Mirvit requires the usual 10-15min more so that rmadison agrees13:43
Mirvah, I don't know if it has changed recently13:44
sil2100I remember it always being accurate since some time13:45
sil2100But hm, dunno13:45
pmcgowanMirv, all set for that webkit landing? I just wanted to be sure it got properly tested13:47
Mirvpmcgowan: oSoMoN is reviewing the webbrowser branch still (or maybe done now), I can start builds and testing as soon as traincon-0 is over.13:50
oSoMoNMirv, webbrowser-app branch reviewed and approved13:51
MirvoSoMoN: thanks!13:52
pmcgowanMirv, whats the webbrowser branch for?13:52
ogra_=== image triggered ===13:54
sil2100o/13:55
sil2100ogra_: thanks!13:55
ogra_bot shoulld announce it soon13:55
mhr3sil2100, reconf of 010 pls?13:57
sil2100mhr3: done13:58
mhr3ty13:59
sil2100np13:59
imgbot=== trainguard: IMAGE 63 building (started: 20140603 14:00) ===14:00
ogra_there we go14:01
ogra_with luck we should have the first smoke results when the meeting starts14:01
Mirvpmcgowan: https://code.launchpad.net/~dbarth/webbrowser-app/webapp-container.remove-qtwebkit-support/+merge/221773 <- oSoMoN can probably elaborate (or dbarth). I'm not 100% sure if it'd be required, since we do continue to have qtwebkit (just a newer version).14:03
pmcgowanMirv, we need some sort of test plan to validate that webkit is still working properly independent of our not using it anymore14:03
pmcgowanMirv, we didnt include it originally for a reason, to do with bad dpr, not sure thats an issue or not14:04
Mirvpmcgowan: I think the DPR got fixed, the only issue we had with 5.2 (as we were going to ship it originally) at the end was that WebGL sites didn't work properly with it14:05
Mirvbut yes, I added now a tab to create a test plan at https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AjuCdq68GSyVdFI4QzNQdWpfME5aMEV2VXo0cUpOMkE#gid=2314:06
pmcgowanMirv, thanks14:08
sil2100ogra_: as with every Tuesday, I might not be able to attend the meeting14:09
ogra_ok14:09
sil2100ogra_: but I will be back around 20 if anything14:09
Mirvdbarth: can you check ^ url (it's editable). I'd mostly want at least one line for every theoretical webkit user, and status if it's something that is needed to be retested with 5.2 or has already switched to oxide14:09
ogra_yeah, thats fine14:09
* sil2100 is battling the bot14:09
dbarthMirv: yup14:09
ogra_sil2100, since the AP tests wont be done anyway we should probably skip ... unless people feel an urge to talk14:10
dbarthMirv: there is an additional apparmor change required to do a transparent switch though14:10
ogra_i assume we'll be somewhere in the middle14:10
dbarthtrying to set that up properly14:10
Mirvthanks, it's starting to look good. I wasn't sure if OA switched, but indeed it did.14:13
pmcgowanMirv, can you look at https://bugs.launchpad.net/qt/+bug/1318584, I cannot seem to assign it to you14:47
ubot5Ubuntu bug 1318584 in Next Generation Checkbox (GUI) "checkbox-gui crashed when switching video out mode to external only mode" [Medium,Confirmed]14:47
mhr3sil2100, could you click the rebuild button on https://launchpad.net/~ci-train-ppa-service/+archive/landing-010/+build/6065421 pls?14:56
mhr3i was too quick to build it, the ppc -api pkg wasn't published yet14:56
mhr3Mirv, or you ^?14:57
sil2100mhr3: after my meeting14:57
sil2100mhr3: click15:00
mhr3thx15:01
=== retoaded changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp - sprinting this week, slow to respond | CI Train support - US: robru, rsalveti - EU: sil2100, Mirv | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: TRAINCON-0; QA Lab VPN connection issues
mterrykgunn, Saviq: how do I initiate a new silo request?  Ask here?15:07
Saviqmterry, added to the spreadsheet?15:07
mterrySaviq, no not yet, I'm new to "adding new silos to spreadsheet" -- wasn't sure of order of operations15:07
Saviqmterry, so just add your data at the end of https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AuDk72Lpx8U5dFVHQ3FuMDJGLUZCamJfSjYzbWh3Wnc#gid=015:07
Saviqmterry, now you want to set the "Ready" column to Yes15:16
Saviqmterry, and start begging sil2100 to get us the last silo available ;)15:16
mterrySaviq, well...  I'll have to poke tiagosh about getting the checklist right for his MP15:16
mterryIf I'm going by book15:16
Saviqmterry, we'll probably have to wait until image 63 results are in, so there's time15:17
sil2100;p15:17
sil2100I like the description!15:17
Saviqsil2100, see!15:18
sil2100Saviq, mterry: assigned! Silo 020 for you15:19
Saviqsil2100, we'll be quick, promise15:19
mterryheh, that's a bit confusing but makes sense since it was just cleared by us15:19
Saviqindeed ;)15:19
mterrySaviq, tiagosh is going to add a checklist to his merge15:19
Saviqmterry, sure, we can build in the mean time15:20
Saviqmterry, as that doesn't affect the process15:20
mterrySaviq, yup, just saying15:20
Saviqkk15:20
Saviqmterry, you pushing the button on this?15:20
mterrySaviq, the build button?  I guess the fact that landing-020 sheet has content means its configured with those 3 branches already?15:21
Saviqmterry, when sil2100 says it's ready, that's when it's ready ;)15:23
sil2100hm, should be safe to do operations15:23
sil2100But the spreadsheet is behaving strangely15:23
sil2100Could anyone of you press the build button?15:24
imgbot=== trainguard: IMAGE 63 DONE (finished: 20140603 15:25) ===15:25
sil2100o/15:25
imgbot=== changelog: http://people.canonical.com/~ogra/touch-image-stats/63.changes ===15:26
ogra_yippie !15:26
sil2100Image testing begiiin!15:26
sil2100\o\ /o/15:26
ogra_now the anxious waiting til the smoke infra picks it up15:26
popeyupdated to #63 and i have no wifi ☹15:29
sil2100Reboot15:30
sil2100Oh15:30
sil2100On the other hand15:30
popeybah15:30
popeyjust did15:30
sil2100My landing might have kicked us in the balls15:31
* ogra_ watches the update animation idly 15:31
sil2100popey: still no internet?15:31
popeyworks after reboot15:31
sil2100Ah, phew15:32
sil2100Still, let's see if ogra_ will have the same15:32
ogra_wlan oon flo is fine here15:32
ogra_right after upgrade ... no extra reboot15:32
ogra_mtp also behaves15:33
sil2100Indicators present?15:33
ogra_yep15:33
popeyyes15:33
ogra_all like it should here15:33
sil2100Ossum15:34
ogra_hmm, going back after adding a U1 account is weird15:35
ogra_you see system settings for s split second ... then it slides away and shows you systme settings15:35
ogra_on first sight it looks all fine15:37
ogra_so lets hope the AP tests confirm that15:37
ogra_popey, do you feel the need to have a landing meeting now ? would be only us two and i think we know what we wait for now to drop traincon-015:52
popeyi do not15:52
ogra_great, then lets skip15:53
ogra_and just watch smoke tests to chug along15:53
ogra_mediaplayer had no failures !15:53
ogra_that hasnt happened in a while15:53
=== retoaded changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp - sprinting this week, slow to respond | CI Train support - US: robru, rsalveti - EU: sil2100, Mirv | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: TRAINCON-0
pmcgowanogra_, 63 looks better then16:04
pmcgowannice job cross fingers16:04
ogra_pmcgowan, from a user POV it looks ok16:04
pmcgowanhah16:04
pmcgowanimplying not from an engineers I suppose16:04
ogra_cant say much about the AP tests yet16:04
plarssil2100: no landing meeting?16:04
ogra_plars, we decided to skip ... all we wait for is anyway that the AP tests finish16:05
ogra_plars, to drop TRAINCON-016:05
ogra_(i didnt know you are back today)16:06
plarsah16:06
plarsogra_: I am16:06
ogra_ok16:06
plarsogra_: fyi, I'm gone all next week though16:06
ogra_well, we hpefull have all fixes proper now16:06
ogra_so waiting for #63 to finish testing16:06
ogra_if that looks okayish we'll drop TRAINCON-016:07
ogra_plars, why does mako look so much different while the tests are running ... i.e. the tablets both list all tests, mako seems to be missing 1/316:07
plarsogra_: so I didn't really look into or understand the fix for this yesterday, but I do see that dbus-x11 is still installed. is that expected?16:08
ogra_it isnt installed16:08
plarsogra_: could be part haven't started yet, I'll take a look16:08
plarsogra_: it is installed, I'm looking at it right now on 6316:08
ogra_http://people.canonical.com/~ogra/touch-image-stats/61.changes16:08
plarsroot@ubuntu-phablet:~# apt-cache policy dbus-x1116:08
plarsdbus-x11:16:08
plars  Installed: 1.6.18-0ubuntu616:08
ogra_i definitely dont have it on a fresh flash on 6316:09
plarsthat's running in the lab right now16:09
ogra_ugh16:09
plarsso something is bringing it in then16:09
ogra_thats evil16:09
plarsogra_: that's what I was trying to say last night16:09
popeyyeah, my #63 has it16:09
plarsSo I don't expect 63 to look any more green16:09
ogra_root@ubuntu-phablet:~# dpkg -l |grep dbus-x1116:10
ogra_root@ubuntu-phablet:~#16:10
ogra_root@ubuntu-phablet:~# system-image-cli -i|grep "version version"16:10
ogra_version version: 6316:10
ogra_how can you have it ... i dont get that16:10
ogra_popey, ?16:10
ogra_did you install any AP tests ?16:11
popeynope16:11
popeyoh hang on16:11
popeyit's possibly I'm an idiot16:11
popey  Installed: (none)16:11
ogra_PHEW !16:11
popeysorry ☻16:11
plarsah, found it16:11
popeycarry on taking the pills16:11
plarsmessaging-app-autopilot brought it in16:11
ogra_plars, well, if that got pulled in by any AP this is super evil16:12
ogra_it will cause all sorts of dbus flakyness16:12
ogra_i hipe the autoremove at least purges it after testing ?16:12
plarsogra_: http://paste.ubuntu.com/7581296/16:12
ogra_there is a fixed dbus-x11 too, that shouldnt spawn any dbuses anymore16:12
ogra_oh16:13
ogra_ofono-phonesim16:13
plarswill find out, that's the one its on now16:13
ogra_i think ofono-phonesim-autostart depends on it for whatever reason16:13
ogra_if we get dbus-x11 1.6.18-0ubuntu8  in that shouldnt be an issue anymore though16:14
ogra_but i see you are on ubuntu6 in the lab16:14
ogra_so lets cross fingers16:14
ogra_anyway, the trigger that made it start random dbus-daemons is gone so it shouldnt break16:15
ogra_we re-ordered the greeter session startup so that you have a dbus before anything  wants to talk to it16:15
plarsogra_: well, we don't apt-get update16:15
ogra_which hopefully makes dbus-x11 less harmful16:16
ogra_i dont think ubunt8 has migrated yet anyway16:16
ogra_likely still in proposed ... it has a gazillion of deps for which it runs all tests16:16
ogra_https://launchpad.net/ubuntu/utopic/+source/dbus/1.6.18-0ubuntu816:17
ogra_hah16:17
ogra_Published 13 minutes ago16:17
ogra_bad timing16:17
plarsogra_: autoremove didn't take it out :(16:18
ogra_crap16:18
ogra_i wonder fi thats not responsible for flaky tests we saw in the past then16:18
plarsogra_: quite possibly16:18
ogra_well, messaging-app at least passed fine16:18
ogra_ARGH!16:26
ogra_dialer-app failed all tests16:26
ogra_:((((16:27
robruogra_, also, no indicators on #6216:28
ogra_robru, 62 is dead and buried16:28
ogra_it was horribly broken16:28
robruogra_, is there a new one? I just flashed and got 62...16:29
ogra_<ogra_> root@ubuntu-phablet:~# system-image-cli -i|grep "version version"16:29
ogra_<ogra_> version version: 6316:29
robruogra_,16:29
robruogra_, ha!16:30
ogra_robru, 63 is fine from locally running tests and from a few mins usage testing16:30
ogra_but it seems dbus-x11 is still causing issues in the lab16:30
ogra_sigh calendar is also failing16:33
ogra_mterry, ^^^16:33
ogra_http://ci.ubuntu.com/smokeng/utopic/touch/mako/63:20140603.1:20140530/8381/16:33
ogra_plars, for a test, can you re-run any of the failing ones after manually purging dbus-x11 again ?16:34
ogra_we definitely had them pass in local tests16:34
robruogra_, so what happened in 63? did that have the greeter revert?16:38
ogra_nope16:38
ogra_that would have been a gigantic revert16:38
ogra_we didnt want to do that16:38
ogra_we fixed it instead16:38
robruogra_, yeah, i don't want that either ;-)16:38
robrugreat16:38
ogra_apparently not enough though16:38
robruah16:38
ogra_looking at the currently running AP tests16:39
ogra_but thats seemingly caused by a leftover dbus-x11 package that should not be on the device at all16:39
plarsogra_: sure, of course. I could do it locally right now, or I could do it in the lab after the runs are done16:40
ogra_plars, in the lab please, i want to be sure you use the same env and setup16:40
ogra_plars, we need to find a way to purge dbus-x11 though ... i dont get why teardown doesnt pull it out16:41
plarsogra_: 'sudo apt-get autoremove --purge -y messaging-app-autopilot' was what got run, I suppose maybe because it didn't follow the dep chain far enough down16:42
ogra_well, autoremove should actually follow the chain16:43
ogra_plars, could we somehow ass dbus-x11 as a permanent entry to the teardown packageset ?16:53
ogra_*add16:54
plars:)16:54
ogra_heh16:54
plarsprobably, let me take a look - it would be a temporary hack I guess right?16:54
plarsogra_: is it even worth it though? from what you said, in the version of dbus-x11 we pull in next it shouldn't have this problem16:54
ogra_well, we definitely dont want that package ever installed ... except for ofono-phonesim tests ...16:54
ogra_well, the new version will only make sure that the Xsession script wont run if there is already a dbus running16:55
ogra_but if you have any code that calls gdbus this will spawn a session dbus via the dbus-launch binary which dbus-x11 ships16:56
ogra_in which case you can end up with multiple dbuses again16:56
ogra_which will get you flaky tests most likely16:56
ogra_so it might or might not fix it in the lab16:57
ogra_it will fix one case we know about16:57
ogra_but there can be more ...16:58
ogra_in whicch case it is safer to always purge it16:59
plarsogra_, doanac: provided I didn't do something stupid, https://code.launchpad.net/~pwlars/ubuntu-test-cases/kill-dbus-x11/+merge/221925 ought to take care of it. Have to force that to pass in case we ever end a test where it didn't get installed, which will certainly happen17:03
plarsogra_: when I get an ack on it I'll merge it and this branch will get used to test after 63 is done. since dialer app failed I'll include it, which should also bring in dbus-x11 and we'll make sure it gets removed properly17:04
ogra_plars, looks ok to me, did you try it on a local device ...17:04
ogra_not sue about the quoting now that you have multiple packages17:05
plarsnot yet, about to. I just need to dig out my phone17:05
plarsactually...17:05
plarsI may have a better way17:05
ogra_our issue is that we need to get out of TRAINCON-0 ... so even if you can re-run the tests with purging it manually after dialer-app and messaging-app have run would help a lot17:06
ogra_we need some reliable data, things are piling up17:06
plarsogra_: we'll get there, I'm testing that branch now and once 63 finishes I'll rerun things with this17:09
ogra_awesome17:09
plarsit has my full attention right now17:09
ogra_heh, thanks :)17:10
plarswell, except for part of my lunch sitting on my desk here17:10
ogra_right, eat first, we dont want you to fall over starving, then we dont have anyone over there anymore :)17:10
plarsogra_: hah, you just saw me last week, I would hardly describe me as "starving" :)17:10
ogra_haha17:11
ogra_we all gained some wieght over the years17:11
ogra_a *little*17:11
=== alan_g is now known as alan_g|EOD
mterryogra_, sorry was out to lunch.  63 is expected to have some rough edges.  silo 020 should fix them?  (i.e. work with dbus-x11)17:51
Saviqhuh, so we didn't get any improvement on 63 re: smoke testing...17:51
Saviqah, probably because it still has unity from yesterday...17:56
Saviqor well, Friday, actually17:56
* sil2100 back18:12
sil2100ogra_: do we have results?18:12
josharensoncihelp, I need a parameter added to mir-mediumtests-runner-mako18:18
sil2100ogra_: 63 doesn't look any better...18:18
plarssil2100: we're working on it, turns out dbus-x11 gets installed with some of the AP tests, but autoremove doesn't clean them18:19
sil2100phew18:19
doanacjosharenson: what sort of parameter are you needing?18:19
sil2100plars: thanks! So there's still chance we might not have to revert? ;)18:19
josharensonI need "mir_performance_tests" added to the test_suite section18:19
plarssil2100: I just pushed a patch to force the removal, but trying to get complete results on mako first. I just noticed that one of the makos never installed though because it never went into fastboot. We seem to have a new problem...18:19
plarsogra_: ^18:19
josharensonChrisGagnon and I added it last week, but it looks like it never stuck18:20
plarsogra_: sometimes when I do 'adb reboot-bootloader', it looks like it just reboots the device back into the OS18:20
plarsso it never actually goes to fastboot, and it just gets stuck trying to run ubuntu-device-flash18:20
doanacjosharenson: i'll have to check with fginther about that. i'm not sure the exact mechanics involved for that18:20
josharensondoanac, ok18:21
fgintherjosharenson, doanac, hello. I can help with that18:25
plarssil2100: I've restarted the failed ones (except gallery since we know that's a different problem) on mako, this time it should remove dbus-x11 by force18:26
fgintherjosharenson, it sounds like all you need is to add another test to the list of test suites?18:26
josharensonfginther, correct18:26
sil2100plars: \o/18:29
sil2100plars: thank you, you're our hero!18:29
plarssil2100: :( I much prefer to play the villain. I must work harder on this handlebar mustache18:31
sil2100hah!18:31
sil2100;)18:31
mterrysil2100, I built silo 020, and it seemed to have worked (packages in PPA), but the status of the build job is aborted?18:31
fgintherjosharenson, doanac, https://code.launchpad.net/~fginther/cupstream2distro-config/mir-perf-tests/+merge/22193918:31
fgintherjosharenson, can you make sure that added test name is correct?18:32
sil2100mterry: hm, let me take a look18:32
sil2100huh18:32
sil2100mterry: ok, now this creeps me out - it has been aborted by 'anonymous'18:32
sil2100I don't remember we had a user like that, maybe that's some sabotage?18:33
josharensonfginther, approved MP18:33
sil2100mterry: if you re-run the build job with 'watch only' checked, it will finish what it started18:33
mterrysil2100, ok18:33
sil2100mterry: i.e. finish with success18:33
sil2100:)18:33
fgintherjosharenson, I'm leaving my wifi for a while. When I get back online, I'll deploy the job updates18:40
ogra_plars, where does the dash in "adb reboot-bootloader" come from ?18:40
ogra_(back btw... had some dinner)18:40
josharensonfginther, sure thing. i'm about to go to lunch, but when I get back I'll monitor new builds.18:40
josharensonfginther just ping me on irc when you deploy. Thanks for the help.18:41
plarsogra_: we have to do that in our provisioning script because ubuntu-device-flash does not18:41
ogra_plars, yes, but the command should be adb reboot bootloader without dash18:41
plarsogra_: oh? I think both might be valid. That's always worked for us18:42
ogra_hmm, i only know the no-dash version :)18:42
plarsogra_: yeah, both of them are in the help18:42
ogra_could be that maguro required a dash18:42
ogra_ah, k18:42
plarsogra_: I have one phone where the problem is pretty reproducible, let me play with it a bit18:43
plarsperhaps they are slightly different and the no-dash one is better now18:44
plarsnope18:44
plarsstill getting right back to device mode18:44
ogra_yeah, its a flag that gets handed to the kernel who hands it to the bootloader ... from there it ends up on the cmdline as value of bootreason=18:44
plarsI wonder if it's just this device... I've seen other devices work just fine18:44
ogra_chekc the cmdline18:44
plarsbootreason=watchdog18:44
ogra_hmm, never seen that18:44
plarsogra_: I'm starting to fear maybe this device is doing bad18:45
ogra_well, perhaps the bootloader is18:45
plarsogra_: yeah, I took this device offline for now and I'll see if it shows up on any others18:46
ogra_try re-flashing it with plain android ... also check the rest pf the cmdline and compare it with a working one18:46
ogra_androidboot.bootloader= should be interesting18:46
plarslast successful flash on this phone was with 6118:47
plarsso it's worked recently18:47
ogra_shoudl show you the version18:47
ogra_right, but if the last android that was flasshed had a different bootloader version that could be related18:47
ogra_so you might hit a bug with that ... we never touch bootloader or radio firmware18:48
ogra_lol .. the watchdog logs are funny ...18:50
* ogra_ just looks into some 18:50
ogra_[335722.400536] Watchdog bark! Now = 335722.40051918:50
ogra_[335722.400908] Watchdog last pet at 335711.09292018:50
ogra_[335722.401116] cpu alive mask from last pet 0-318:50
ogra_[335722.401314] Causing a watchdog bite!18:50
ogra_*woof*18:50
bfillerrobru: trying to use citrain host-install 20 script and getting these errors: http://pastebin.ubuntu.com/7582208/18:51
bfillerrobru: any ideas?18:51
bfillerrobru: actually is that the right command? want to install silo 20 on my device18:52
* sil2100 prepares some dinner18:52
bfillernm18:53
robrubfiller, hehe, no you want device-install. or probably device-upgrade for a unity silo18:53
* ogra_ goes afk for a bit again ... hopefully we have greener tests when i return :)18:53
bfillerrobru: ok right, got confused because when I ran citrain-slurp it said deprecated to use citrain host-install18:54
robrubfiller, yeah... slurp and host-install are the same... ;-)18:54
robrubfiller, you would have wanted citrain-push with the old tools18:54
bfillerrobru: ok got it, thanks18:54
robrubfiller, you're welcome!18:55
robrubfiller, although that error is odd anyway... it should have just installed those things on your computer, not sure why it wouldn't find the package.s18:55
plarsogra_: interesting, I'm talking to sciri now, looks like we have 3 makos that appeared to be stuck at the google screen but they were booted. All of them are behaving this way18:58
* ToyKeeper wonders why apport keeps popping up and using half the CPU, with only like 10 seconds between instances.19:11
sil2100plars, ogra_, robru: so far results seem really good!19:15
plarssil2100: we haven't started to get to the interesting ones yet19:17
plarssil2100: but I suspect this will make things improve quite a bit19:17
ogra_sil2100, patience19:25
ogra_i said the same thing last night19:25
ogra_plars, well, take a close look at the bootloader version etc ... beyond that ... i would blame mmc wearout, take a look at syslog/dmesg on them ... its not like these phones get used "normal" :)19:27
=== ted is now known as tedg
plarsyeah, look for the 97 failures to start dropping before getting excited. That will be a bit longer19:51
ogra_yup19:52
tedgHmm, why are silos rebuilding? Are we decided on the greeter in/out thing?19:54
tedgOr is the bot just going crazy.19:54
ogra_tedg, i would suspect the latter19:57
ogra_the greeter in/out thing will be decided once we have complete test results19:57
ogra_then the channel topic will also be updated19:57
ogra_(it is looking promising  fwiw)19:58
tedgogra_, okay, thanks. Didn't realize that was in the topic, it was ellipsized for me.19:59
ogra_yeah, topics arent a good place ... but easy to point to ;)20:01
ogra_now there comes an exciting one ...20:02
* ogra_ sees clock-app running 20:03
ogra_and passed !!!!20:05
ogra_oooh ... thats looking so promising !!!20:08
ogra_i terminal app failure ... we had that before ...20:08
sil2100:D20:09
* sil2100 writes the e-mail20:17
sil2100I guess the results look far better than before, so the chances of us going down from TRAINCON-0 are highish20:18
ToyKeeperHello jet lag...  how are you today?20:23
plarssil2100: ok, weather is running now, after it completes we should actually see the failures start to drop quickly20:24
sil2100plars: \o/20:25
plarsit appears to be passing now, and was completely failing before20:25
plarssil2100: down to 77 now20:31
plarsfilemanager will pass too20:32
sil2100plars: so far it looks very similar to what we had before20:32
sil2100So I'm really really happy!20:32
plarssil2100: when will the next image build?20:32
sil2100plars: the cronjob I think is around 3 UTC I guess?20:37
plarssil2100: ok, since this experiment seems successful and mako is the one getting the most attention, are you ok to leave things as-is on flo/manta and let them run on the next image?20:39
plarssil2100: mako will have a complete set of results with the workaround for the dbus-x11 removal in 63 though20:39
sil2100plars: so the fix is only for mako for now, right?20:43
ogra_looking good...20:43
sil2100plars: what will we need to do to get all others working?20:43
plarssil2100: no, it's generic enough. I can rerun them if you really need, will just require some babysitting on my part20:43
ogra_nah20:44
ogra_we'll get a new image by cron is a few hours20:44
sil2100Nah20:44
sil2100Just asked out of curiosity20:44
ogra_should be enough to have proper data tomorrow20:44
sil2100As long as other images will be ok20:44
ogra_sil2100, first thing we land tomorrow should be the gallery-app fix20:44
sil2100ogra_: right, I think we can land it even now20:44
ogra_now that the rest looks so nice that red really sticks out :)20:45
sil2100I mean, we wait until we get full results, we drop TRAINCON and BLAM, UITK lands!20:45
ogra_i would like to see what dialer app does20:45
sil2100Then BOOM and the image is broken20:45
ogra_that behaved very odd20:45
sil2100BOOOOOM AND BROKEN20:45
ogra_and sadly uses dbus-x1120:45
sil2100Sorry, got carried away20:45
ogra_haha20:45
sil2100btw.20:50
sil2100ogra_, robru, Mirv: we might consider moving the landing-related discussions to -release20:50
ogra_you mean in general ?20:51
sil2100As slangasek mentioned, it's a release related thing, so it should be discussed somewhere more public20:51
sil2100Yeah20:51
ogra_ok20:51
ogra_no prob with that20:51
sil2100I know we were worried about spamming channels with our landing 'bla bla', but well... slangasek gave us the blessing!20:51
ogra_prepare for a lot of critics from ScottK then :)20:52
sil2100:|20:52
sil2100He didn't vote for my MOTU application!20:52
sil2100Yet!20:52
ogra_he is a grumpy nitpicker ... but usually has a point20:52
ogra_but while he is a pita he also kind of is our community conscience ... it is good to get reminded soemtimes that we dont forget about community20:55
sil2100Indeed21:01
sil2100But he's a bit rude I noticed, this is my only complaint about him21:02
cjwatsonI think this basically a result of feeling largely ignored21:02
ogra_so how about i do something ... like ...21:02
cjwatson*this is21:03
=== ogra_ changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp - sprinting this week, slow to respond | CI Train support - US: robru, rsalveti - EU: sil2100, Mirv | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: -
sil2100ogra_: ++21:03
ogra_weheee :)21:03
sil2100Sending out the e-mail :)21:03
ogra_(i removed TRAINCON-0 )21:03
cjwatson(IOW, he feels he has to be, um, "direct" in order to get anywhere at all)21:03
ogra_in case someone misse dit21:03
ogra_cjwatson, yeah21:03
sil2100robru: ^ and as per e-mail, we're off from TRAINCON-021:04
ogra_sil2100, so do you want to let the gallery-app fix in now so we get it in the cronned image ?21:04
sil2100Yes :)21:04
sil2100Pressing teh button21:04
ogra_or if rob works today we can leave it to him :)21:04
ogra_yay21:04
sil2100(have some connectivity problems though)21:05
* ogra_ is so happ ywe finally got that sorted21:05
ogra_that were two painful days and nights21:05
ToyKeeperHmm, 22 KiB/s from Click's servers to my phone.21:06
ToyKeeperEr, unity8 probably shouldn't be sustaining 50% to 60% cpu load while the phone is idle with the screen off.21:06
sil2100I published something, hope I published the right UITK landing ;p21:06
ogra_ToyKeeper, it also shouldnt consume twice the amount of ram :)21:07
ogra_ToyKeeper, what kept us in TRAINCON-0 and caused so many issues was a pretty giant landing ... watch out for more regressions there are surely some nobody has found yet21:08
ogra_its like an easteregg hunt :)21:08
ToyKeeperSeems that the app scope's app preview screen uses 50% to 70% of the CPU, regardless of whether the screen is on.21:08
ogra_definitely worth a bug21:08
ToyKeeperI think I'm going to be up rather late tonight filing all the bugs I found today.21:09
ogra_well, i hope you test with 6321:09
ToyKeeperThis was on #63.21:09
ToyKeeperMuch better than 62, BTW.  :)21:09
ogra_everything between 59 and 63 was konwn to have very serious issues21:10
ogra_yeah, still not the final fixes, mterry is currently replacing the workarounds21:10
ogra_but surely also wants to know about all regressions21:10
ogra_*but he21:11
* ogra_ cant type anymore after two 16h days 21:11
* ToyKeeper facepalm... just realized I flashed #63 over the notes for the bugs I found in #62 this morning21:11
ogra_and on that note ...21:11
ogra_oh my21:12
ToyKeeperYeah, sounds like bed time.21:12
ToyKeeperI'm jetlagged and barely conscious, but nowhere near done with today's tasks.21:12
robruhey i'm around, was just on lunch. what needs landing?21:17
robrusil2100,  ^21:17
sil2100robru: landed UITK just now21:17
robrusil2100, what about gallery in silo 11?21:17
robrusil2100, also do you know what's going on with the spreadsheet? seems the bot is going crazy pinging things21:18
sil2100robru: it's not tested21:18
robrusil2100, true.21:19
sil2100robru: well, the spreadsheet does do some crazy stuff today21:19
sil2100Ok, it's late now, time to get a life!21:20
sil2100:D21:20
sil2100robru: don't worry about the bot, the spreadsheet did break for me just now, so it can also be related...21:21
sil2100robru: keep an eye on the spreadsheet reverting itself21:21
sil2100You can find the backups here:21:22
sil2100http://sil2100.vexillium.org/citrain-backups/21:22
slangasekogra_, sil2100: I don't think ScottK would object to having more of the discussions around things that are landing happening in the "traditional" channels; my guidance is that people should in general be less afraid of offending someone by talking in the "wrong" channel21:23
tedgHmm, it's late, but would test results for silo 18 be valid now?21:23
tedgDo I need a rebuild?21:23
ogra_slangasek, i didnt mean we would annoy him, i said sil2100 should be prepared for more critics in a channel where he is around (and will see the landing process or even occasionally look at patches)21:25
robrutedg, when was the last build?21:25
tedgrobru, Sunday21:25
ogra_slangasek, i aslo dont think thats a bad thing ... but he can be quite a nitpicker at times :)21:26
slangasekogra_: ah, you mean more community oversight ;-)21:26
ogra_right :)21:26
tedgrobru, I believe you said there was a way to test a silo with upgrade only?21:27
robrutedg, yeah, 'citrain device-upgrade 18'21:27
robrutedg, I just checked your silo contents, of all the packages in your silo, I don't see any new releases for any of them since may 26th, so I guess you don't need a rebuild (it seems all the latest panic fixes didn't touch any pckages in your silo)21:28
tedgrobru, Great! I'll retest, but looking at the greeter changes I'm not too worried.21:28
tedgTrying to get kenvandine off my back, he's heavy!21:29
tedg;-)21:30
robruhmmmm, hang on a sec21:30
robrutedg, I was checking the dates in the version numbers, which would be build dates, not publish dates...21:30
robrushould check again21:30
tedgHow does one get publish dates?21:30
robrutedg, http://people.canonical.com/~rbpark/citrain/ click on the package names, it links to the lp source package pages, it shows version numbers and date uploaded21:31
robrutedg, so ubuntu-touch-session was uploaded on the 1st... but according to the build log that's in there. so now I'm doubly sure your silo is safe to test ;-)21:32
tedgHeh, great!21:33
slangasekogra_: regarding your latest mail on the smoke tests - could I talk you into writing something on https://wiki.ubuntu.com/LandingTeam/IncidentReports to explain how the test envs were different / why ?21:33
tedgReflashing first though.21:33
ogra_slangasek, sure, but not tonight anymore :)21:33
slangasekogra_: ack - thanks21:34
ogra_slangasek, i think it is still not clear to us why apt-get autoremove --purge didnt remove the package alongside though ... the workaround i wrote about is simply adding dbus-x11 to that apt command21:35
josharensonfginther, looks like the new test is running and generating artifacts just fine. Thanks.21:45
slangasekogra_: alright21:52
cjwatsontedg: if it's expired from robru's citrain pages, then https://launchpad.net/ubuntu/+source/SOURCE-PACKAGE/+publishinghistory is a good thing to have smart-bookmarked/memorised21:54
cjwatson(for things that have been published to the primary archive and are no longer in silos)21:55
tedgcjwatson, Ah, cool. Thanks!21:59
robrucjwatson, hmmmm, suddenly I want to create a package that contains a ton of smart-bookmarks for all these little things (i have dozens...)22:02
robruogra_, hey, what's going on in silo 20? if the fix already landed, then what's this silo doing just being "built" without being published yet?22:04
ogra_robru, the real fix hasnt landed ... its hacks all the way down ... 20 has the proper fixes22:05
robruogra_, ooooooh, ok thanks.22:05
ogra_:)22:05
robrumterry, Saviq you guys need any help testing silo 20?22:06
ogra_talk to mterry and Saviq for details22:06
ogra_heh, snap :=22:06
ogra_:)22:06
robru;-)22:06
sil2100o/22:20
tedgrobru, sil2100, testing looks good.22:44
tedgrobru, How do you want to handle landing 18?22:44
tedgrobru, I'm going to EOD. Bear in mind that a publish requires a NEW queue ack.22:56
tedgNot sure if that means DO IT so it can get in the queue, or wait :-)22:57
robrutedg, well, probably DO IT22:57
tedgDO IT!22:57
robrutedg, because everything else will just get stuck in -proposed anyway22:57
robru;-)22:57
* tedg invokes peer pressure22:57
robrutedg, well we're out of TRAINCON-0, so there's no reason I can't. and I know it wont' get through NEW before the next image builds, so I think it's fine22:58
robru;-)22:58
tedgHeh, okay.22:58
robrutedg, https://ci-train.ubuntu.com/job/landing-018-2-publish/29/console ;-)23:00

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