/srv/irclogs.ubuntu.com/2013/11/13/#ubuntu-ci-eng.txt

* Mirv manages to work on Qt still as jenkins doesn't respond06:07
Mirvdidrocks: FYI I'm continuing with Qt related work for now, good progress over there at least while no news about CI yet07:18
didrocksMirv: yeah, I was trying to fetch news :)07:21
didrocksMirv: what are you doing on Qt btw?07:21
didrocks(just curious ;))07:21
Mirvdidrocks: the 5.2beta1 packages I made earlier had some problems with docs&co, but I spotted the problem now and trying to get more proper packages built. lots of modules not updated yet, too, and then getting QtC + plugin sponsored to trusty now.07:22
didrocksexcellent!07:23
didrocksplugin as upstream Qt plugins?07:23
MirvI'm also happy that I've somewhat updated build machine (haswell CPU), eases up07:23
didrocksupstream QtC plugin?07:23
Mirvdidrocks: Qt Creator Ubuntu plugin07:23
Mirvdidrocks: not upstreamed, unfortunately :(07:23
didrocksMirv: well, this one will work through dailies, right?07:23
Mirvdidrocks: it'd work, yes, but I find that some tests would be needed. although now we're in manual mode anyhow so in that sense the daily build could be enabled too.07:24
Mirvthere's currently no tests at all for QtC + plugin functionality07:24
didrocksMirv: yeah, please enable daily build07:25
didrocksas we are in manual mode07:25
Mirvdidrocks: ok.07:27
Mirvit'll need this QtC update as well since it only builds against newer QtC, but now it's therefore a correct time to remove the disabling of daily build from config07:29
didrocksyeah07:31
vilahi guys, still a bit sick but back nonetheless08:14
sil2100o08:14
sil2100o/08:14
vilaI managed to brick my phone yesterday08:15
didrocksvila: hey!08:15
viladidrocks: o/08:15
didrocksmorning sil210008:15
vilasil2100: o/08:15
vilait first failed to download image #20 several times08:15
sil2100Morning!08:15
vilathen succeeded in downloading image #17 (wow, how come ?)08:15
vilaand installing that led to the 'ggogle' logo at boot and stayed there08:16
vilabefore I attempt phablet-flash again, is there something I need to know ?08:16
didrocksvila: system update?08:17
viladidrocks: yeah, bricked while using OTA08:18
vilasystem update08:18
didrocksvila: ok, ogra knows more what happened, system update was broken yesterday: https://bugs.launchpad.net/ubuntu/+source/system-image/+bug/125018108:18
ubot5Ubuntu bug 1250181 in system-image (Ubuntu) "Duplicate files in winning path should prevent updates" [Critical,Confirmed]08:18
didrocksthen, I saw that stgraber discussed about it08:18
didrocksand remove some files, regenerate some signatures08:18
didrocksI guess that's why you downgraded to image 17 ;)08:18
didrockscan you keep it broken for the meeting?08:19
didrocksso that we can discuss it08:19
vilasure08:19
vilait's not as if it was my main phone... err wait08:19
didrocks:p08:19
vilaha yeah, looking at the bug, I saw many ... interesting numbers in the progress bar ;) The "oops, far more than 100, let's go back so more reasonable number" was entertaining08:20
vila*to more08:20
didrocksvila: I can just testify it's not the UI :)08:21
* sil2100 is saddened by cu2d not being accessible still08:25
sil2100At least appmenu-qt5 will be happy ;)08:25
didrockssil2100: how is it going? something we can test for you?08:30
viladidrocks: also, images are still created despite ci being down ?08:38
didrocksvila: we did some, but without test results, there is little use08:45
popeyvila: it's easily fixed, by downloading the boot file and manually pushing it to the phone when in recovery mode via fastboot08:45
vilapopey: what boot file are you talking about ?08:49
popeytrusty-preinstalled-boot-armhf+mako.img08:50
vilapopey: will that wipe the data ?08:50
popeyno08:50
vilapopey: and where is that ? It seems I can't find it on system-image.ubuntu.com08:52
popeyhttp://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/20131107.1/trusty-preinstalled-boot-armhf+mako.img08:53
vilapopey: thanks08:53
popeynp08:54
vilapopey: can't 'adb reboot fastboot', 'ad reboot recovery' is ok but fastboot seem to be stucked at the google logo08:58
vilapopey: i.e. fastboot boot <file> stays at '< waiting for device >'08:59
popeyhmm09:04
vilapopey: hmm, may be I'm not using the right commands >-/09:05
popeyhmm, does adb push work?09:05
popeyi thought i used fastboot but it's gone from my history09:05
popeyah no, i still have the historty09:05
popeyfastboot flash boot trusty-preinstalled-boot-armhf+mako.img09:05
popeyvila: ^09:06
vilapopey: right, I was missing 'flash', trying09:06
vilasame, waiting for device09:06
vilapopey: may be I need the device to be in the "right" state ? Which state is that and how do I get there :)09:07
popeyadb reboot-bootloader09:08
popeyfastboot flash boot trusty-preinstalled-boot-armhf+mako.img09:08
popeyis what I did I think09:08
vilapopey: works far better, and now 'Start'09:09
vilahmm, no better, still stucked at the google logo09:10
popeyhm09:12
ogra_vila, when was that, that you got r17  ?09:13
vilahold on !09:13
ogra_this should be fixed since tonight09:13
vilaogra_: yesterday09:13
popeyogra_: so I should be able to OTA update mine?09:13
vilapopey: just needed to be a bit patient, it finally booted09:13
ogra_popey, yes, it should give you r2009:13
popey307MB!09:14
ogra_vila, ah, then i'm fine ...09:14
ogra_popey, yes, the screwup on the server actually enforces a full image upgrade09:14
popeyah okay, thats cool09:14
popeyi love this whole OTA thing09:14
vilapopey, ogra_ : So, now that it booted, I'm at image #1709:15
popeyadb shell system-image-cli -i09:15
popeywill tell you09:15
vilaogra_: any explanation for the bricked behaviour after the upgrade to #17 ?09:15
popeyvila: known bug09:15
vilapopey: good (well sort of ;)09:15
popeyvila: bug 125018109:15
ubot5bug 1250181 in system-image (Ubuntu) "Duplicate files in winning path should prevent updates" [Critical,Confirmed] https://launchpad.net/bugs/125018109:15
ogra_vila, the same as for 16,18 and 19 ... broken udev upload09:16
popeyoh, well, thats related but not the cause09:16
vilaogra_: so upgrading to #20 is fine ?09:16
ogra_and that you even got r17 was the bug popey points to09:16
* popey updates to #2009:16
vilaack09:16
ogra_we had actually backed it out, but this bug kind of screwed that up09:16
vilaogra_: reformulating to ensure I understand: #17 was buggy and shouldn't have been proposed as an update but the bug above screwed that and I got it nevertheless and bricked my device09:19
ogra_vila, right09:20
ogra_(16, 18 and 19 still had the same bug, you should have gotten #20 from yesterday afternoon on ... )09:21
didrocksI guess the question is why you got proposed image 17 today09:21
didrocksand not #2009:21
ogra_he didnt09:21
popeyyesterday09:21
didrocksah, ok09:21
ogra_thats why i asked when he got it as my first question ;)09:21
vilayesterday, I got #17 proposed yesterday after having #20 proposed (see my comment on bug #1250181)09:22
ubot5bug 1250181 in system-image (Ubuntu) "Duplicate files in winning path should prevent updates" [Critical,Confirmed] https://launchpad.net/bugs/125018109:22
ogra_looks like you upgraded in the middle of stephane fixing the server09:22
vilabeing almost offline (in my bed toying with the phone the few times I was awake ;) I hoped something has been fixed server side and that I was fine upgrading to #1709:23
ogra_(jumping from 20 to 17 kind of indicates that he was actively shuffling images around at that time)09:23
vilawhen the phone got stucked at the google logo, I just fall asleep again09:23
popeyinterestingly 20 on my phone is running but I just have a google logo09:24
vilawith full confidence the issue will be sorted out when I'll wake up ;)09:24
popeybut adb shell shows unity8 is running09:24
ogra_runs fine for me09:24
vilapopey: that's the new screen saver maybe ?09:24
popeyhah09:24
popeyoh, it's not rebooted properly09:24
popey 09:25:01 up 1 day,  7:47,  1 user,  load average: 1.32, 1.32, 1.3409:25
popeyalso.. I have two phones and I'm adb shell'ed into the wrong one09:25
* popey starts again.09:25
popey /ignore popey09:25
ogra_oh, i like the new click scope09:25
* popey makes a cup of tea for the call09:26
* ogra_ goes for fresh coffee, brb09:26
vilaogra_, popey: using adb shell s-i-c seems to find #20, going that route09:48
ogra_vila, note that on the bug too please09:48
vilaogra_: on it09:49
ogra_thx09:49
ogra_=== Image #21 building ===09:49
popey\o/09:56
vilalol, just got #20 installed, rebooting09:58
vilaogra_, popey: there shouldn't be any difference in upgrade paths between UI and system-image-cli right ?09:58
ogra_the UI only attaches to s-i-c afaik09:58
vilaogra_: ack, hopefully09:59
vilaanyway, I've put the s-i-c output in the bug report, not sure the UI uses exactly that or some d-bus interface but they may find a difference10:00
vilaotherwise, it's not the first time the UI is capricious for me which may be caused by some local network glitch (dunno why the -cli won't run into that though)10:00
vilaupgraded to image #21 via UI, all went well11:16
didrocksvila: the UI uses the d-bus interface of s-i-c11:18
viladidrocks: ack.11:22
=== gatox is now known as gatox_brb
=== gatox_brb is now known as gatox
ogra_whoops11:57
ogra_=== Image #21 DONE ===11:57
ogra_(since ages, sorry)11:57
popeyhehe11:58
=== MacSlow is now known as MacSlow|lunch
MirvI was wondering too :)12:03
=== alan_g is now known as alan_g|afk
asacdidrocks: you changed the time of landing standup from next tue?12:21
asac(just saying in case you wanted to change it starting today)12:21
didrocksasac: I only changed it during UDS12:21
asacic12:22
didrocksremoved the afternoon one and started it a little bit before UDS12:22
asackk12:22
asacmakes sense12:22
=== alan_g|afk is now known as alan_g
=== alan_g is now known as alan_g|lunch
=== MacSlow|lunch is now known as MacSlow
asacev: didrocks: this charting tool you use, is that also available for other chart types?13:25
=== alan_g|lunch is now known as alan_g
fginthermorning13:57
didrocksasac: no, this one is only a very light online javascript thing for this kind of diagrams14:22
asacthx14:22
=== greyback is now known as greyback|away
=== alan_g is now known as alan_g|tea
=== alan_g|tea is now known as alan_g
=== greyback|away is now known as greyback
elopiocihelp, can the landing bot reject an MP if it doesn't have a given number of approvals by the members of a team?16:20
sergiusenselopio, I would guess that's a feature request16:20
cjohnstonIt seems like it wouldn't reject it, just that it wouldn't accept it16:21
cjohnstonthough I don't know what it actually does in practice16:21
sergiusenscjohnston, it uses lanchpad lib to get the approver list; once you have that, it shouldn't be hard16:22
sergiusensprobably even enforce certain review types16:22
elopiowith reject, I meant: move from approved to needs fixing.16:22
cjohnstonI get that, but why would it be moved to needs fixing?16:22
sergiusensI am not a fan of that; whoever happroves has a certain responsibility there16:23
cjohnstonthis seems like a social problem that is trying to be fixed with code16:23
sergiusenscjohnston, +116:23
cjohnstoncan it be setup to require 3 reviews, yes its possible (I don't know if it is available currently, but it would be possible to add if it isnt)16:23
sergiusensif my MR is one line of code that just fixes a README I would certainly not want 10+ reviewers16:23
sergiusenswe'd end up bundling MRs all in one16:24
elopioI would like that we have a strict policy that all the code be reviewed at least by one member of the team.16:24
cjohnstonI believe it should already be setup that way16:24
elopioon the tarmac of u1, that's checked when the branch is marked as approved.16:24
elopioand it's moved to needs fixing so you can tell the bot when the branch is ready again by approving.16:25
cjohnstonI believe that MPs currently require a proper team member to approve it16:25
elopiosergiusens: I think that even one-liners should need one review. There are typos in a big number of the cases.16:26
elopiobut 10, of course not. I'm talking of enforcing one review.16:26
sergiusenselopio, sure, but it's variable16:26
cjohnstonagain, its a social problem.. if the person who marks the top as approved doesn't ensure that there has been an approved review then there is another problem16:26
sergiusenselopio, most of the type the person doing the top approve is the one doing the review16:26
elopiocjohnston: yes, sometimes it's the same person that writes the code the one that top approves. That's mostly when we are in a hurry, but it happens.16:27
sergiusenselopio, so people just do that16:27
cjohnstonelopio: 1) social problem 2) that person should still ensure that a review has been done16:27
elopioso, I agree that we can just talk to the teams to avoid that, and we will do it.16:27
sergiusenselopio, not necessarily; sometimes it's an autoland gone bad (infra issues) and the same author just doing what he's already been green lighted to do16:27
cjohnstonas long as it doesn't merge without a review, isn't that what is needed?16:28
elopioit would be nice to have jenkins making sure that they follow the policy, so we don't have to keep an eye on every project to check that.16:28
cjohnstonelopio: I believe it does require a review16:28
cjohnstonby a person on the proper team16:28
sergiusenselopio, cjohnston what is the policy? Nothing prevents you from just merging manually into trunk either16:28
cjohnstonsergiusens: I believe that jenkins requires one approval16:29
sergiusenscjohnston, not really; if autolanding passes it just changes the vote for ps-jenkins to be of Approved16:29
cjohnstonis ps-jenkins a member of the proper team?16:29
sergiusenscjohnston, yeah, ps-jenkins is a member of everything under upstream merger16:30
elopiosergiusens: yes, you are right, you can always merge manually.16:30
cjohnstonwe are in a call right now, but possibly fginther could provide a little more insight on the setup16:30
elopiobut if we agree that we need at least one review, and never to merge manually into trunk, then what the jenkins bot would be doing is just a reminder that you are missing one review.16:31
cjohnstonit would make more sense to have a comment rather than mark it as needs review16:32
cjohnstonor needs fixing or whatever it is16:32
elopiocjohnston: yes, I agree to that. But then how will you tell jenkins to pick the branch again when you got the review?16:32
cjohnstonagain though, this is trying to solve a social problem with code16:33
cjohnstonI'd be interested in knowing how often this problem happens?16:33
elopiocjohnston: agree to that too :) We'll try to solve the social problem talking with the team managers to follow the policy. But it would be nice to have an automatic reminder when you don't follow it.16:34
cjohnstonelopio: who is doing the happrove though?16:34
elopiocjohnston: I can't give you numbers. For what I've seen, it's not common, but it tends to increase when we are in running, which can makes things worse.16:37
elopiothe happrove is done by a member of the team, not necessarily different from the committer.16:37
ogra_didrocks, i'm wondering, do we actually need a meeting today ? seems CI is still mostly idling16:37
ogra_(nothing exciting about images either)16:38
cjohnstonok, so if they are happroving why aren't they reviewing (if it isn't the comitter) or if it is the comitter, why are they happroving without reviews?16:38
didrocksogra_: I think we don't16:40
ogra_yeah16:40
didrockssil2100: vila: fginther: ev: cyphermox: kenvandine: as long as CI isn't up, I think we don't need a meeting anymore, we can skip that one ^16:40
fgintherdidrocks, +116:40
didrocks(would be nice to have a status on the infra though)16:40
kenvandine+116:40
ev+216:41
sergiusenselopio, cjohnston fwiw this topic comes back every now and then16:41
fgintherdidrocks, I don't have any status yet16:41
cjohnstonthere was supposed to be an email coming out soon16:41
cjohnstonre: status16:41
cyphermoxok16:41
evI'll provide a status shortly16:41
* ogra_ goes back to play with his new laptoy16:41
elopiocjohnston: I haven't seen a happrove from <not-the-committer> without a review, but it can happen.16:41
elopiocjohnston: I have seen a self happrove with no reviews when the change is considered trivial or too important to wait for a review.16:41
fgintherelopio, I believe there is a bug already for your original request (or something very similar016:41
evI'm just waiting for Larry to provide me a few more details so it's less "we have 80% of the racks cabled" and more "we're having m-o back up at foo PM Boston time"16:42
fgintherelopio, will look it up in a moment16:42
elopiothanks fginther.16:43
didrocksogra_: oh! new laptop :)16:43
ogra_yeah16:44
kenvandinex86 even!16:45
sil2100didrocks: ok, less interuptions for me then :)16:46
sil2100ev: awesome16:46
elopiosergiusens: that's why we are trying to define the details we expect from our projects16:47
elopioit's here, in case somebody is courious: https://docs.google.com/a/canonical.com/document/d/1D3G62wd1wMAH9zoHiHYXkabnHqelqz8Upffi0d_6iaA/edit#heading=h.nzp18cbqfdgy16:47
elopiocomments are still welcome.16:47
=== gatox is now known as gatox_lunch
popeydidrocks: landing call?17:01
didrocksoh sorry, didn't ping popey :)17:01
didrockspopey: as long as CI isn't up, I think we don't need a meeting anymore, we can skip that one17:01
popeykk17:01
popeythanks17:01
* didrocks pitti phew, 2 hours of discussion to push hard for dep8 in the way CI should describe the tests17:01
didrocks:)17:01
didrockswe'll have a happy Martin!17:01
=== gatox_lunch is now known as gatox
=== alan_g is now known as alan_g|EOD
fgintherelopio, I finally found the bug report: https://bugs.launchpad.net/jenkins-launchpad-plugin/+bug/113443518:27
ubot5Ubuntu bug 1134435 in jenkins-launchpad-plugin "Allow for only limited folk to HApprove MRs" [High,New]18:27
fgintherelopio, it's specifically for rules around top approve, but it's a similar concept.18:27
fgintherelopio, So the short answer is that it is not supported now, but we're planning to build it into the CI Airline system18:28
elopiothanks fginther, that will do for.18:28
=== fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: use 'fginther' for assistance | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: All services are down (1SS move)
=== fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: use 'cihelp' for assistance | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: All services are down (1SS move)

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