=== robru changed the topic of #ubuntu-ci-eng to: Train trouble? ping trainguards | CI problems? Use JenkaaS: http://bit.ly/jenkins-docs | Train: http://bit.ly/1hGZsfS | QA Signoffs: http://bit.ly/1qMAKYd | Known issues: sil2100 away on Friday, Mirv sick leave but occasionally around, robru sprinting in Texas [04:59] dobey: cjwatson: kenvandine: yeah so it was not just because train packages don't tend to have debian/source, but the train refused to build a source tarball (for some reason) when it was kept, so it was not able to upload the package to PPA. that was the problem before getting to the build/install problems. [05:00] dobey: cjwatson: kenvandine: the first problem might easily be caused by changes in train, and then finally with your finding we'd have the explanation for the second problem.. === Mirv changed the topic of #ubuntu-ci-eng to: Train trouble? ping trainguards | CI problems? Use JenkaaS: http://bit.ly/jenkins-docs | Train: http://bit.ly/1hGZsfS | QA Signoffs: http://bit.ly/1qMAKYd | Known issues: sil2100 away, Mirv sick leave but occasionally around, robru sprinting in Texas [06:24] the sick leave part was technically correct yesterday but it seems unfortunately that it will hold true today again [07:15] Mirv: ping [07:22] robru: ping === Trevinho_ is now known as Trevinho [08:46] Mirv: the "for some reason" really needs to be investigated in more detail IMO; having debian/source/format should be considered standard practice nowadays [08:47] jibel, I forgot today is a holiday here and in the UK, probably makes sense to move our meeting to tomorrow [08:51] Saviq, are you off by one day? yesterday was an uk holiday afaik but not today [08:51] seb128, I might be, there is a holiday here, though ;) [08:51] jibel, ↑ [08:53] Saviq, np, next week then [08:53] Saviq, I'll be off Thursday and Friday and I cannot do it tomorrow [08:53] jibel, ack [08:53] rhuddie, ^ [08:53] sry === vrruiz_ is now known as rvr [09:23] cjwatson: it's probably the creation of the orig tarball that has changed that doesn't anymore like .bzrignore/.gitignore files that are lower than / of the source tree: https://ci-train.ubuntu.com/job/ubuntu-landing-058-1-build/54/console [09:23] so it's mostly ubuntu-push specific special case [09:23] it first excludes them and then notices they are "local changes" [09:23] I'll ping robert when I'm both back and robert is also available [09:23] Mirv: Ah, yeah, that does seem like a plausible kind of ci-train bug [09:33] Mirv, hey, what's the latest? :) I see the quilt 3.0 patch didn't fix it? [09:39] jgdx: sorry I'm on sick leave today and have about 1h now in which I need to clean stuff and eat, I can't work on that. check with morphis he did some build(s). from what I've quickly viewed if the quilt still gives problems to train (like yesterday), the solution is to keep the removal of it, .install files as unchanged and add chmod +x debian/*.install in debian/rules [09:40] Mirv: thanks, I think we're close to get this fixed [09:40] will work with jgdx to finish this thign :-) [09:50] jibel, hey, on silo 69. it failed the auto signoff because of this: https://requests.ci-train.ubuntu.com/static/britney/vivid/landing-069/excuses.html [09:51] jibel, it's the same tests that pass on ubuntu-ui-toolkit but not on ubuntu-ui-toolkit-gles. bzoltan said this happens frequently for uitk. [09:51] jibel, however, in order to rerun this, we'd need pitti to trigger it because of some bug in the system and I can't reach him atm [09:51] jibel, tl;dr: IMO you could start with QA-ing this [09:52] I'll try to get autopkgtests retriggered asap [09:54] mzanetti, okay [09:57] Mirv: FWIW I believe the ci-train bug is in tools/buildpackage.sh, which does --exclude='.bzr*' --exclude='.git*' when building the orig.tar [09:58] it'll need to do that only at the top level somehow [09:59] or if that's too fiddly (I'm not quite sure how to do it ...), switch to just --exclude=.bzr --exclude=.git and accept that ignore files will end up in the tarball, which wouldn't be terrible [10:20] morphis, ^ my bad, kicked off another [10:21] jgdx: ok [10:23] jgdx: updated https://requests.ci-train.ubuntu.com/#/ticket/1361 [10:23] so it has test plan etc. [10:23] thx [10:40] morphis, argh, that one flaky test failed ^ we have a patch for it, but it's a rather impactful change [10:40] jgdx: hm, can we just disable it? [10:41] morphis, maybe, but let's try another build first? [10:41] whatever you want, really, your silo :p [10:41] jgdx: hahaha :-) [10:42] I'm going to prepare a branch disabling it while we try again, ok? [11:18] morphis, https://code.launchpad.net/~jonas-drange/ubuntu-push/lp1577723-skiptest/+merge/293606 [11:18] ok [11:20] though seems the http13 tests have started failing [11:24] wonderful [11:33] jgdx: sounds like all those tests are not very stable [11:33] morphis, those tests are just copy pasted from the golang go-http repo [11:34] why that? [11:34] morphis, wish I knew [11:34] :-) [11:35] jgdx: who developed ubuntu-push intially? [11:36] I think it just always existed [11:38] It was John Lenton and Samuele Pedroni [11:38] AIUI [11:38] Only goes back to 2014! Not that old :) [11:41] :-) [11:41] jgdx: I would say for the current situation we can disable them and bring them back afterwards [11:41] otherwise we just loose more and more time [11:42] jibel: ^^ [11:42] morphis, what do they test? [11:42] jgdx: ^^ [11:42] pedronis | Chipaca ^^ [11:43] hopefully it only goes back to 2014 so they won't have anyone else to ping :) [11:44] jibel, they prob test the go http client library because they use it heavily. And it did uncover a bug in go itself some time back [11:45] why idk [11:45] jgdx, I don't know the details but it should be tested in go itself, push seesm like the wrong place. But there is probably a good reason [11:46] jgdx, anyway the change is still just state -> State? [11:46] ++ [11:46] jibel, some packaging fixes and disabling a test which will require a fix that we cannot include for OTA11 [11:46] jibel, see bug 1577723 [11:46] bug 1577723 in ubuntu-push (Ubuntu) "Postal message order is wrong" [Undecided,New] https://launchpad.net/bugs/1577723 [11:47] jgdx, can you skip this test and move on [11:48] jibel: afaik the http tests are now failing [11:49] morphis, all of them? [11:49] jgdx can comment on that [11:55] well, this run was very green, one failure on armhf, not related to http [11:56] all green on vivid, one singular test failure on xenial armhf [11:57] * ogra_ glares at the last ubuntu-pd image build failure mail ... [11:57] why the heck is it installing ltsp-server by default ?!? [11:59] oh, thats even in the builder chroot before it calls live-build [11:59] ChrisTownsend: ping [11:59] rvr: Hi [12:00] ChrisTownsend: Hi. I'm testing silo 31 [12:00] ChrisTownsend: It crashes when the german layout is selected [12:01] ChrisTownsend: At least, the app opens and closes [12:01] rvr: What crashes? The X app? And strange since that is one of keyboard layouts I tried:) [12:02] ChrisTownsend: I first tested Spanish and it worked fine. Then added German, and didn't work. [12:02] ChrisTownsend: I thought it was because two layouts were selected, but I removed Spanish. And then German. [12:02] ChrisTownsend: If no layout is selected, apps open again fine. [12:02] rvr: Are all X apps affected by this? [12:03] ChrisTownsend: gedit and LibreOffice [12:03] Those are the ones I'm using to check [12:03] rvr: So you select the keyboard first in System Settings, put German layout on top, and then try to start the X app, right? [12:05] ChrisTownsend: Yes [12:05] rvr: Ok, I'll try to reproduce when I get to my office where my test phone is. Thanks for letting me know. [12:06] ChrisTownsend: Apps closes when I select another layout and go back to it [12:07] rvr: You are using the Puritine click too, right? [12:07] rvr: Could you also look in ~/.cache/upstart/application-legacy-puritine_gedit_0.0.log and see if anything of note is in it? [12:08] rvr: If you have time. I can try to look later. [12:10] rvr, okay, we have testable vpn servers for almost all of the auth schemes [12:11] rvr, pete-woods has instructions for testing cert+pass, and cert-only will be done against canonical's company vpn [12:11] jgdx: Wee! [12:11] ChrisTownsend: Hmmm === alan_g is now known as alan_g|lunch [12:11] ChrisTownsend: I'm using latest rc-proposed image [12:12] rvr, note there is no way of testing Static key auth right now, so we're hoping QA can be of help here (either compromising testing or help us find some server that let us test it) [12:13] rvr: I guess I was meaning how are you getting the X apps on the device? They are not in rc-proposed by default. [12:13] ChrisTownsend: They are in frieza [12:13] rvr: Oh Frieza, nevermind [12:15] rvr: I'm still in the N4 mindset:) [12:15] ChrisTownsend: http://paste.ubuntu.com/16200922/ [12:16] rvr: Ok, got it. Definitely an Xmir crash. I'll look into today. Thanks! === _salem is now known as salem_ === alan_g|lunch is now known as alan_g [13:05] jgdx, approved [13:06] thx jibel [13:21] jibel, jgdx: awesome! [13:21] morphis, thanks for your help, that was a tough one [13:24] morphis, once it's published we'll rebuild an image [13:24] jibel: can't wait for that to happen :-) [13:24] jibel: indeed, but thankfully we got it solved :-) [13:27] morphis, jgdx you can publish or need a train guard? [13:28] jibel, I don't know how. [13:28] jibel: its not a main package, right? [13:29] morphis: it doesn't matter because it's only going to the PPA [13:29] anyone can publish [13:31] ok [13:32] jibel, jgdx: https://ci-train.ubuntu.com/job/ubuntu-landing-055-2-publish/10/ [13:34] nice [13:39] jibel, morphis, what about the livecd-rootfs bits (they are still sitting on my disk) [13:39] ogra_, can you upload the changes? [13:40] jibel: afaik ogra_ wanted a build with his previous password db changes first [13:40] jibel, sure, i was just waiting for a "go" from someone [13:40] well, since there debian/ changes, you probably need kenvandine or someone to publish with an ack on the pkg changes [13:40] morphis, well, the NM dropping wont cause any build issues so it can land ... [13:41] ogra_: OK [13:41] (you will at most have runtime issues) [13:42] :-) [13:42] lets hope we don't get htem === fginther` is now known as fginther [14:05] ondra, ToyKeeper, any news on that adb crash? [14:07] charles hi [14:07] sorry jus saw it [14:07] charles we had day off here [14:07] alesage_, i just replied to your comment on the trello card for silo 59 [14:07] charles what devices is crashing on? [14:07] alesage_, i think your problem there was the same issue you had with my silo, but shouldn't be an issue now since we've had images built since then [14:08] alesage_, so should be unblocked [14:08] ubuntu-qa, Hey guys silo 9 is ready to review: https://requests.ci-train.ubuntu.com/#/ticket/1362, and it is target to ota11 [14:09] rvr, hey, what's the status of https://trello.com/c/9ufZ0bTi/3124-1314-ubuntu-landing-008-ubuntu-settings-components-jgdx ? [14:09] ondra, [04:05:37] charles: It was krillin rc-proposed 321 and arale rc-proposed 313. [14:09] ondra, [04:07:14] -*- ToyKeeper tries it again with 323+315. [14:09] ondra, I got the crash only once, on krillin rc-proposed 323 [14:10] jgdx: Still need the VPN instructions :) [14:10] pete-woods, can you abide? ^ :) [14:10] ogra_, ubuntu-push is in the overlay. can you build an image? [14:11] pete-woods, provide rather [14:11] jibel, once livecd-rootfs is published [14:11] jgdx: okay, will start yanking the stuff out of the .ovpn file [14:13] thx [14:15] charles and any steps to reproduce or just to plug cable? [14:20] kenvandine, ack === alesage_ is now known as alesage [15:06] jgdx: I've put the split files in here: http://people.canonical.com/~pete/vpn/vpn-credentials.tar.gz [15:07] user/pass = testuser/testpassword [15:10] grmpf ... [15:10] slangasek, did you change the crontab on nusakan ? [15:10] * ogra_ notes the vivid snappy builds are gone from it [15:13] jibel, your image build is running [15:16] jgdx, jibel pointed out that automated tests are failed for silo 8, do you know what’s up with that? [15:16] pete-woods, is that also for rvr ? [15:16] oSoMoN, checking [15:17] renatu, what’s the status of the new calendar-app release you’re preparing? [15:17] jgdx whoever needs to test it [15:17] jgdx I haven't had to to check it works at this end [15:17] but this is the same thing I did last time [15:17] you have to poke through the config file and map the entries to settings in the UI [15:18] and obviously set the various certificates and user/pass [15:21] ogra_, thanks [15:21] oSoMoN, jibel: failures are in unity8 tests that should be unrelated to my changes. Any way to re-run them? [15:22] jgdx, if you go to the excuses page there is an icon to re-run the tests [15:23] jibel, i'm not allowed [15:23] jgdx, https://requests.ci-train.ubuntu.com/static/britney/xenial/landing-008/excuses.html next to amd64 [15:23] cimi, could you rerun the failing autopkg tests in silo 8? [15:23] the recycle symbol [15:24] oSoMoN, is on silo, I am testing it today [15:24] cimi, direct link: https://autopkgtest.ubuntu.com/request.cgi?release=xenial&arch=amd64&package=unity8&trigger=ubuntu-settings-components%2F0.7%2B16.04.20160427-0ubuntu1&ppa=ci-train-ppa-service%2Fstable-phone-overlay&ppa=ci-train-ppa-service%2Flanding-008 [15:24] I need to test if it is compatible with ota10, to launch it without framework bump [15:25] oSoMoN, https://requests.ci-train.ubuntu.com/#/ticket/1362 [15:41] jgdx, I can't, I'll ask saviq tomorrow or dednick [15:42] jgdx: anyone who has upload rights to the package that failed, or the package that triggered the tests (or who is a coredev) can re-run the tests === chihchun is now known as chihchun_afk === nuclearbob is now known as nuclearbob_afk [16:12] ogra_: I did, by request of jibel [16:13] bregma: ping [16:16] slangasek, you mangled the snappy entries... [16:45] rvr, yo? [16:46] bregma: Hi [16:47] bregma: libertine-scope... it's not showin gedit icon. Also, it's not translated. [16:48] rvr, translation is not a surprise, but it should probably be showing an icon [16:48] rvr, are you testing with Puritine or with your own hand-made container? [16:48] bregma: Latest rc-proposed on frieza [17:00] slangasek: https://code.launchpad.net/~robru/cupstream2distro/fix-ubuntu-push/+merge/293643 [17:21] cimi, okay, np [17:22] bregma: https://trello-attachments.s3.amazonaws.com/5726002b34077f16dbadb21b/1920x1200/025826ed80f3356796d3e1e1188dfce5/libertine-scope-gedit.png [17:29] rvr, https://bugs.launchpad.net/libertine-scope/+bug/1558741 is in progress, should hit OTA-12 [17:29] Ubuntu bug 1558741 in Libertine Scope "Not all icons are picked up when surfacing apps" [Undecided,New] [17:30] maybe OTA-11 if things aren't too sluggish === charles_ is now known as charles [18:13] ondra, ToyKeeper, sorry my ISP went out for a couple of hours there. Either of you still around? [18:16] charles hi [18:17] ondra, any news on the adb crash? Is there anything you need from me for that silo? [18:29] charles I need to know how to reproduce it, do you have some bug report? [18:29] steps [18:29] alesage: "sim services" can be grayed out if your sim card doesn't have service numbers, not related to the changes proposed though [18:29] charles and it it crashing only with your silo? [18:29] charles code handling those requests is from google, we might be up for some AOSP contribution :) [18:36] ondra, I got it to crash by walking through the manual test indicator-display/adb-prompt-after-boot, and then following the steps ToyKeeper wrote in comment #1 at https://trello.com/c/gzf3hdwh/3111-1299-ubuntu-landing-035-indicator-display-charles-ondra [18:37] according to ToyKeeper it is only crashing with the silo, but I didn't test w/o the silo [18:38] let's see... [18:38] ogra_: I commented out a single ubuntu-touch line, AFAIK; what was mangled? [18:39] Snappy daily pointed to yakkety again and the vivid line was completely gone [18:40] BTW, is the importer off ? I built an image for jibel ages ago, seems cdimage has them sine a while [18:44] ondra, getting a clean device to test with [18:50] jibel, are we building rootfs yet? :) [18:55] jgdx, long done [18:56] but seems system-image didn't import it yet.... [18:57] http://cdimage.ubuntu.com/ubuntu-touch/vivid/daily-preinstalled/current/ has them [18:57] slangasek, are there known importer issues ? ^^^ [19:04] ogra_: not that I'm aware of currently [19:04] ogra_: and no, I did not touch those snappy lines [19:04] Weird [19:05] I changed them on Friday... must have been someone else who changed them then [19:13] slangasek, hmmmhttp://paste.ubuntu.com/16207648/ [19:14] bah...http://paste.ubuntu.com/16207648/ [19:15] something is clearly broken [19:21] ogra_, do you know what is this empty custom tarball? [19:21] which channel/device? [19:21] ogra_, they changed to their own jenkaas instance maybe something broke there [19:22] jibel, sorry, no idea... and it is ages since I have even seen the system-image code last [19:23] the tarball seems to be empty though [19:23] if I can belive the message [19:24] ogra_, yeah maybe the importer (or whatever copies images from the system that builds the tarballs) as not been reconfigured to point to the new location [19:24] has* [19:24] and the file in the old location has been zeroed ? [19:25] it should just pull in the old file if thats the case [19:26] just a blind guess, the way custom tarball are imported is pretty obscure to me. But this new jenkaas to build custom tarballs is one thing that changed recently [20:24] robru: I just realized this never landed... is there any chance you might be able to help? https://code.launchpad.net/~toykeeper/phablet-tools/phablet-network-no-nm/+merge/251564 [20:25] ToyKeeper: what do you need? is it not in a silo? [20:25] robru: The MP got approved but apparently never merged into trunk. [20:25] robru: ... and you're listed as top contributor to that project. :) [20:30] ToyKeeper: reviewed [20:31] ToyKeeper: lol that's a year old [20:32] ToyKeeper: wait, disregard my review [20:33] Yeah, I had forgotten about it; thought it landed because it kept working after I dist-upgraded. [20:34] ToyKeeper: so anyway, this needs to go through a silo like anything else. you want to create one? [20:35] robru: Sure, though I'm not sure how that works from the dev side. [20:35] ToyKeeper: log in at requests.ci-train.ubuntu.com and clicke 'create new ticket' and then fill it out as best you can [20:36] robru: Thanks, will do. Wasn't sure if there was a documented process to follow or if I should just figure it out. :) [20:37] ToyKeeper: mountains of verbose and questionable recency documentation is at https://wiki.ubuntu.com/citrain/LandingProcess, but it should be pretty straightforward to just fill out the form. [20:37] :) === blr_ is now known as blr [20:58] ToyKeeper: click build [20:58] ToyKeeper: also you might want to consider making this a yakkety silo, as it's probably aimed at desktop users? [21:00] does phablet-tools actually land via train? [21:01] ah i guess so. guess i'm just used to having it from the sdk team ppa [21:01] dobey: yeah typically I do a train silo then copy it to that ppa. [21:01] since I can't be bothered to build packages locally [21:02] robru: Yes, desktop users, mostly the ones in QA, so the distro QA is using (and everything after that). [21:02] ToyKeeper: I dunno what QA is using ;-) [21:03] ToyKeeper: I guess do a xenial+vivid build, I'll copy those to the sdk ppa, then do a yakkety build and that can go to archive. [21:20] aww phablet-tools isn't arch:all === salem_ is now known as _salem [22:42] ogra_: did you look at /srv/system-image.ubuntu.com/www/full/pool/custom-51fb7bd3e49898416dd4d9ce47c71563743aefac760b034834a793fa19d2619a.tar.xz ? it's an html 'openid auth required' page [22:42] ogra_: so do we know which import that came from, that we can disable it? [22:54] ogra_: looks like this came from https://jenkins.canonical.com/ues-phone/job/custom-bq-aquaris.en-mako/lastSuccessfulBuild/artifact/out/here-international/build/custom.tar.xz