=== timchen1` is now known as timchen119 === nudtrobert1 is now known as nudtrobert [05:26] * pitti NBSes the old linux-raspi binaries [06:50] ^ disables apport for the final release and fixes the failing test case === henrix_ is now known as henrix === Laney changed the topic of #ubuntu-release to: Released: Trusty 14.04.3, Vivid 15.04, Wily Beta 2 | Archive: final freeze, britney block in place | Wily Release Coordination. Please don't upload things during freezes where you shouldn't, or be prepared to apologise to the release team | We accept payment in cash, check or beer | melior malum quod cognoscis [09:40] darkxst: If you upload, I'll be happy to review. [09:43] infinity: can you add the netboot tests for rc please [09:44] jibel: ^-- You know how to add all those "fake" tests? [09:45] * xnox lols at "fake" [09:45] infinity: jibel is travelling to get to bluefin, I'll run the tests now and add the results after [09:45] davmor2: Ahh, he's coming here today? I couldn't remember. [09:45] infinity: yeah travelling to the airport now [09:46] infinity: that's why I asked you :P [09:49] flexiondotorg: I'm not sure I understand your ubuntu-mate-meta "fix". If there's a file conflict, fix the packages in question, rather than unseeding them, surely? [10:17] infinity, uploaded [10:19] darkxst: No matching clutter for that mutter? [10:21] darkxst: Oh, I guess they're not really tied together. I'm not braining well today. [10:26] infinity, no, this does not involve clutter [10:26] can do a 0day update for clutter though [10:27] insert ^sru [10:27] jamespage: Around? [10:27] infinity, yes [10:28] jamespage: Any ideas about the nova autopkgtest regression on ppc64el? [10:28] adt-run [09:09:12]: test nova-compute-daemons: [----------------------- [10:28] OK [10:28] ERROR: NOVA-COMPUTE IS NOT RUNNING [10:28] adt-run [09:09:45]: test nova-compute-daemons: -----------------------] [10:28] adt-run [09:09:45]: test nova-compute-daemons: - - - - - - - - - - results - - - - - - - - - - [10:28] nova-compute-daemons FAIL non-zero exit status 1 [10:28] oh joy [10:28] jamespage: Seems pretty consistent over many re-runs. [10:28] lemme look in a bit [10:28] I would suspect a racey test that might pass everywhere else by accident [10:30] infinity, I'll get coreycb to take a look; I have a few things to sweep up before I have capacity today [10:30] jamespage: It's holding back cinder and nova, would be nice if you get your Top Men to give it some analysis. [10:30] infinity, ack [10:32] infinity, is there a nice place to debug that failure? looks like its related to nova-compute-lxc on ppc64el [10:32] jamespage: kelsey01, perhaps? [10:32] jamespage: Depends on if you need root. [10:33] jamespage: If you need root, I might be able to spin you a VM somewhere. [10:33] infinity, I need root really - vm would be good === willcooke_ is now known as willcooke [10:50] jamespage: Which ssh key? [10:51] jamespage: ssh -p 6222 jamespage@10.245.64.15 [10:59] jamespage: *poke, poke* :P [11:08] infinity, Morning. [11:09] infinity, I am happy to fix the Compiz packages but I simply don't have the time this week. [11:09] infinity, sorry - electrician in today as well :-) [11:09] infinity, So I'd like to unseed Compiz for armhf, since it is not essential (in fact useless for the Pi 2) at this point. [11:09] infinity, can I get root on that as well? [11:10] infinity, Can you please update ubuntu-mate-meta for me? [11:10] infinity, I can then build the Raspberry Pi 2 image so it is ready for 22nd Oct release day. [11:11] infinity, And I'll submit fixes for Compiz after 15.10 is out. [11:21] jamespage: You have root, check ~/sudo_password [11:21] infinity, gotit - ta [11:55] infinity, pitti: do the autopkg tests for ppc64el run inside a kvm? or is it container based? [11:57] jamespage: It's lxc in kvm, I believe. [11:58] infinity, this might be the limitation - tests are working fine under just kvm [11:58] jamespage: Except that armhf is also lxc, afaik, so that doesn't add up. [11:58] infinity, yeah - I was just looking for a correlation there [12:11] infinity, Please can you let me know if you will be able to update/upload ubuntu-mate-meta? [12:11] flexiondotorg: I can, yes. I was just questioning the reasoning for it. But it also doesn't matter if you want to drop compiz from armhf, so uploading. [12:12] infinity, Many thanks! [12:12] flexiondotorg: http://paste.ubuntu.com/12859967/ <-- Look right to you? [12:14] infinity, Perfect! [12:18] jamespage: the tests are running in lxc, yes; the host is kvm [12:19] * jamespage scratches his head [12:19] jamespage: given the history on http://autopkgtest.ubuntu.com/packages/n/nova/wily/ppc64el/ the test is obviously racy, but I wondered if 7 failures in a row with the new version is just bad luck or a systematic problem [12:19] pitti, test is OK without LXC [12:20] pitti, I suspect that armhf might have the same problem - but fails slower that ppc64el [12:20] http://autopkgtest.ubuntu.com/packages/n/nova/wily/armhf/ is much more green (although still a bit racy indeed) [12:21] pitti, I have a feeling that set of tests is going to behave like that under lxc [12:21] * pitti runs it manually with shell'ing in [12:21] infinity, Many thanks! [12:22] apt-get install -y nova-compute $daemon 2>&1 > /dev/null [12:22] if pidof -x nova-compute > /dev/null; then [12:23] jamespage: ^ i. e. the postinst doesn't wait until it's actually started? [12:23] pitti, hmm - that might be it [12:24] or it fails to start; not that easy to tell due to the >/dev/null [12:24] jamespage: nova-compute-lxc might just fail under lxc sometimes? [12:25] it might [12:25] pitti, I think I'll drop the install re-direction that might give a clue [12:25] jamespage: on the last 6 logs nova-compute-kvm is consistently working and -lxc is concistently failing [12:25] pitti, yeah - that might be race again [12:26] nova-compute-kvm will actually get installed via tests/control [12:26] so has longer to be present in the process listing [12:26] jamespage: ah, due to it being the preferred alternative [12:26] yeah [12:28] so if nothing fundamentally changed in teh last version, should we hint this so that it can land in the next rebuild? [12:28] I guess this is a bit urgent [12:28] pitti, nova/cinder not on any images [12:28] ah, good [12:28] ssh gloin.kernel [12:28] -EWRONGWINDOW [12:33] jamespage, hey, need anything? [12:34] jamespage: yeah, seems to be a race indeed [12:35] jamespage: after the test fails, nova-compute is running [12:35] (for -lxc) [12:38] # service nova-compute start; echo "STARTED"; pidof -x nova-compute [12:38] STARTED [12:38] (no pid) [12:38] jamespage: ^ [12:39] tjaalton: Why has there been no attention to LP: #1507255 ? [12:39] Launchpad bug 1507255 in xserver-xorg-video-intel "Garbled graphics in Wily with Intel GMA4500 chip" [Undecided,New] https://launchpad.net/bugs/1507255 [12:39] Launchpad bug 1507255 in xserver-xorg-video-intel (Ubuntu) "Garbled graphics in Wily with Intel GMA4500 chip" [Undecided,New] https://launchpad.net/bugs/1507255 [12:39] tjaalton: Looks pretty vile. [12:41] infinity: the bug is one day old, first time I hear about it too [12:41] tjaalton: Oh, fair point. His mention that it's been happening for a while made me not check the date. [12:43] jamespage: so my first gut feeling is that the start job is done when python runs, but that still needs to parse/read/run nova-compute and change its argv etc. [12:44] jamespage: so maybe add a retry loop for 5 s or so? [12:47] coreycb, nah on this [12:48] jamespage: testing a fix now [12:48] coreycb, hey - could you work the changes/uploads for the horizon upgrade issue we discussed on friday [12:48] coreycb, that will need an SRU as well - lets see if we can get that in before zul's stable kilo release work [12:48] jamespage, sure will do [12:49] pitti, a retry? [12:49] jamespage: testing http://paste.ubuntu.com/12860302/ [12:50] pitti, awesome ta [13:12] infinity: re bug 1507333 - the guy who reported that was also talking about corrupt icons - I'll ask him to grab the iso again - ftr I just installed with the same iso and was ok here [13:12] bug 1507333 in systemd "kernel panic during installation" [Medium,Incomplete] https://launchpad.net/bugs/1507333 [13:12] bug 1507333 in systemd (Ubuntu) "kernel panic during installation" [Medium,Incomplete] https://launchpad.net/bugs/1507333 [13:13] flocculant: Corrupt icons too, could point to just dodgy hardware. Which would explain a systemd segv that no one else sees. [13:14] infinity: yep - the guy is in our QA team, so if he'd seen that before I would know [13:15] unless it's broken recently ofc ;) [13:15] anyway - soon as I see him I'll ask him [13:15] infinity, Do I need to request an unblock for ubuntu-mate-meta? [13:15] flexiondotorg: Nah, I'll unblock it. [13:15] Cheers. [13:15] infinity, You in England? [13:16] could someone please review console-setup and localechooser, once that is done we can rebuild d-i and ubiquity to ship the fixes on images [13:16] pitti: Wasn't your last apport upload meant to fix the tests? :P [13:16] flexiondotorg: Yep. [13:16] :-) [13:16] infinity: yeah, and now they fail for something completey different and unexpected; need to investigate [13:17] pitti: \o/ [13:17] infinity: it did fix the test that was failing previously at least [13:18] jamespage, infinity ^ should fix the ppc64el failure; it succeeded twice in a row on wolfe (i. e. on the production platform) [13:18] and also in local (amd64) LXC [13:18] pitti, thankyou [13:18] well, I forgot to take out the extra "echo waiting..", but not a biggie [13:19] if you care, I'll reupload [13:19] it was meant to show me how many 0.1 s iterations it needs and that the loop is working [13:19] * pitti self-rejects and reuploads [13:20] infinity: btw, I uploaded and self-accepted postfix yesterday as that was holding up dpkg; I hope I didn't disrupt any freeze/image building with it [13:20] but I figured we'd want that dpkg everywhere [13:21] pitti: Nope, all good. [13:21] pitti: Thanks for that. === balloons is now known as Guest77565 === Guest77565 is now known as balloons_ === mhall119_ is now known as mhall119 === balloons_ is now known as balloons [14:28] jamespage: Where does "manage.py" in openstack-dashboard-ubuntu-theme.postrm come from? [14:28] jamespage: Pretty sure you can't rely on it being there during purge, only remove, if it's a package that one depends on. [14:28] jamespage: And it's redundant to do it in both cases anyway. [14:29] nova-compute-daemons PASS [14:29] jamespage: ^ http://autopkgtest.ubuntu.com/running.html for nova/ppc64el \o/ [14:30] pitti: Huzzah. Thanks. [14:30] so I guess you might want to lift the freeze for nova and cinder? [14:30] to not ship wily with the release candidates? [14:30] pitti: Already done/doing. [14:31] infinity: ah nevermind, Laney did it already [14:32] infinity, ^^ to fix the virtualbox provide snarfu ... [14:35] pitti: Now if you can either fix apport or tell me I can ignore its broken tests, that would be lovely. :P [14:35] infinity: I hinted it this morning already [14:35] pitti: Oh, so you did. [14:36] infinity: but I'll look at the tests ASAP of course (not today any more though, need to run to French class in 10 mins) [14:38] pitti: why take a french class, just hang around in here and get insulted by jibel, seb128 and cyphermox, you'll be fluent in no time ;) [14:38] tout est cassé ! merde ! [14:38] ;-) [14:39] pitti: I tend to use a different wording, and not so much write that, just say it loud :) [14:39] hahaha [14:39] * pitti n'apprend pas encore à jurer [14:39] (because I'm polite and stuff) [14:40] cyphermox: but I make you swear all the time :P [14:41] davmor2: sure, but I don't swear when other peoples might hear it, because I don't want to subject people to it and sound like I'm always mad :) [14:41] cyphermox: I'm sorry you have another mode???? [14:42] you all seem to forget that next release d-i and ubiquity will default to fr_CA ;) [14:42] ... or fr_QC, depending on how elections end today? ;P [14:43] cyphermox: is that like EN_US ie not quite French :P [14:43] ha ha [14:43] seb128: thanks, your fix for ubiquity works -- I'll just use 50 instead so it's not too large for people with small screens. [14:43] cyphermox, yeah, sorry I couldn't easily test so I picked a number ... thanks for testing/including ;-) [14:44] np. we've been wanting to fix/rework this for a bit, make the detials more useful [14:44] right now it's too small to be very readable, and if we make it larger it's as useless for large screens. [14:44] s/large/small/ [14:46] right [15:01] infinity: weren't we dropping wubi.exe? [15:35] jamespage: Should I worry about this vmware-nsx in the NEW queue that no one has ever bugged me about? [15:36] infinity, I suspect its been there so long its now bitrotted and will FTBFS [15:36] defer it until +1 opens... [15:36] I can always request a backport like I did early last cycle for a few other similar bits and pieces [15:37] jamespage: Happy to reject it if there's no urgency. [15:39] infinity, fine with me === SpamapS_ is now known as SpamapS [16:12] infinity, Are you planning on respinning all the flavours again? [16:14] flexiondotorg: Yeah, we've changed both installers, so $world will respin later tonight to pick all of that up. [16:16] unsure if that's confirmed or just a problem for some users and how much of an issue it is but bug #1507245 and bug #1506502 looks like they are worth investigating a bit [16:16] bug 1507245 in ubiquity "Not all language support files are installed even 'if install updates' is selected during installation" [Undecided,New] https://launchpad.net/bugs/1507245 [16:16] bug 1506502 in ubiquity "fcitx-mozc is not installed by default on Ubuntu 15.10" [Undecided,Confirmed] https://launchpad.net/bugs/1506502 [16:16] bug 1507245 in ubiquity (Ubuntu) "Not all language support files are installed even 'if install updates' is selected during installation" [Undecided,New] https://launchpad.net/bugs/1507245 [16:16] bug 1506502 in ubiquity (Ubuntu) "fcitx-mozc is not installed by default on Ubuntu 15.10" [Undecided,Confirmed] https://launchpad.net/bugs/1506502 [16:16] (trying a japanese install in a vm to see if I can confirm the issue) [16:21] poor ubot93 [16:31] infinity, Thanks. [16:36] bug #1506502 is confirmed [16:36] bug 1506502 in ubiquity "fcitx-mozc is not installed by default on Ubuntu 15.10" [Undecided,Confirmed] https://launchpad.net/bugs/1506502 [16:37] could be because fcitx-mozc is in universe, does the language-selector code grab things from there? [16:38] if not bug #1486772, the package needs to be seeded [16:38] bug 1486772 in gyp "[MIR] mozc" [Undecided,Incomplete] https://launchpad.net/bugs/1486772 [16:38] (the MIR is fix commited for mozc, the status reflect another component) [16:38] seb128: It probably shouldn't. But lemme look at the MIR. [16:41] seb128: It should be seeded in live (and should be moved to main once it is). [17:01] infinity, I've just noticed something weird. [17:01] infinity, mate-sensors-applet is still in wily/proposed. [17:02] infinity, This was synced weeks ago. [17:02] infinity, Can you let me know why it is stuck there and how to move it on? [17:02] flexiondotorg: isn't it on excuses? [17:02] http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html [17:02] * flexiondotorg is looking now [17:03] ^ fails to build on some architectures [17:03] cyphermox, Which ones? [17:04] cyphermox, It build in debian unstable/sid [17:05] cyphermox, Can you request a rebuild? [17:05] Not touching package due to block request by freeze (contact #ubuntu-release if update is needed) [17:06] bad timing ;) [17:06] https://packages.debian.org/sid/mate-sensors-applet [17:06] (that was likely the beta freeze that was in effect when you synced it) [17:07] cyphermox, Can you unblock mate-sensors-applet please [17:07] ogra_, Thanks. [17:10] flexiondotorg: sorry, I can't unblock things, no permissions for that :) [17:10] arf, but it's failing to build anyway [17:10] infinity, Please can you unblock mate-sensors-applet [17:11] OK, can you request a rebuild? [17:11] flexiondotorg: Unblocking won't help. [17:11] I have previously built it in a PPA to test it and it builds in Debian sid. [17:12] flexiondotorg: It has build-deps that can't be satisfied. [17:12] Where can I see this please? [17:12] flexiondotorg: In the build log. === Guest9808 is now known as Adri2000 [17:13] sbuild-build-depends-mate-sensors-applet-dummy : Depends: libxnvctrl-dev but it is not installable or [17:13] nvidia-settings but it is not installable [17:14] infinity, Yep, so I see. [17:14] infinity, Thanks. [17:14] flexiondotorg: In Debian, libxnvctrl-dev seems to exist on all arches, it's only on some in Ubuntu. [17:14] infinity, So I see :-( [17:14] flexiondotorg: So you probably want to arch-restict that build-dep, if it's optional. [17:15] infinity, Incantation to that? [17:15] [amd64 armhf i386] [17:15] But that implies that dropping the build-dep doesn't break the package somehow. [17:15] Which you might want to verify. :P [17:16] infinity, It won't. [17:16] infinity, configure will simply disable nvidia support if the build dep is missing. [17:16] I feel like it might break the new mate-sensors-applet-nvidia package. :P [17:16] infinity, Should I provide a debdiff? [17:16] Or make it so useless that you should also arch-restrict that. [17:17] infinity, Yep, arch restrict that -nvidia package too. [17:17] And the nvidia-dbg package, obviously. [17:18] Ordinarily I'd fix this in Debian and synv to Ubuntu, but no time for that. [17:26] infinity: apw: uploaded a new dfsg rtl8812au, removing those files [17:41] flexiondotorg: So, yeah, gimme a debdiff, I guess. [18:07] infinity: i noticed that the 20151017.1 ISO didn't pass the automated testing... are you expecting to respin the ISO this week, or do you expect 20151017 to be the final? [18:08] jderose: I'm respinning tonight after some bits of the archive settle. [18:08] jderose: Which automated testing didn't pass? [18:09] infinity, Sorry for the delay. Been reading bed time stories to my daughter. [18:09] infinity, How does this look? - http://paste.ubuntu.com/12864333/ [18:12] flexiondotorg: Looks like it might work. Did you test by fudging the arches and doing a local build? [18:12] flexiondotorg: I can try that. [18:12] flexiondotorg: (Also, your changelog indentation is wrong, but I'll fix that) [18:12] Haven't done a local build, just wanted to make sure the ubuntu1 suffix was the right way to do things? [18:13] flexiondotorg: Yup. [18:13] infinity, OK. Do you want me to raise a bug? [18:13] flexiondotorg: No. [18:13] infinity, Thanks! [18:13] flexiondotorg: I'm going to test this a bit here and then sponsor it if it seems sane. [18:14] * flexiondotorg goes off to test PowerPC build for Ubuntu MATE and Lubuntu... [18:16] flexiondotorg: Also missed running update-maintainer(1) (fixed). [18:16] infinity, Thanks. [18:16] flexiondotorg: Testing right now with s/amd64/powerpc/ so my amd64 build will omit those bits. :P [18:17] flexiondotorg: Will upload if this DTRT. [18:17] infinity, Thanks for your help. Let me know if you need me to do anything else. [18:21] infinity: not sure what failed, i just noticed that the migration from pending to current didn't happen, and it was my understand that usually means the automated iso testing failed [18:21] jderose: Fair point. jibel's looking at it. [18:21] infinity: jibel: thanks! :) [18:31] infinity, Thanks ^^^^^^^^^^^^^ [18:48] I have a change to qtubuntu-media that will be published for wily shortly and I want to make sure it gets accepted into the wily release. It's a very minor package dependency change on a specific version of mediascanner (old version now) that mediascanner needs in order to not be blocked in proposed. [19:24] Anyone with thoughts on the qtubuntu-media landing? [19:24] the diff is here: https://code.launchpad.net/~phablet-team/qtubuntu-media/mediascanner-dependency-removal/+merge/274914 [19:28] infinity, Turns out that mate-sensors-applet is busted in Debian too. [19:28] infinity, So I will fix it there too and it can synced for the 1604 cycle :-) [21:05] jhodapp: Laney already fixed it hours ago. [21:33] cyphermox, What options can I pass oem-config to get some debug/logging? [21:40] flexiondotorg: debug-oem-config, I think. lemme look [21:41] yeah, if you pass debug-oem-config on the kernel command-line it will start in debug mode; which is the same as starting it from the console with --debug [21:42] what kind of information are you looking for? === utlemming_sprint is now known as utlemming [21:50] infinity, yeah he forgot to mention that :) [21:50] infinity, but it's only a partial fix [22:11] So has anyone asked sabdfl for the XX name yet, considering last release? :P [22:13] flexiondotorg: so have you hit some issue with oem-config? please share, inquiring minds at system76 would love to know :) [22:25] jderose, Nothing for you to worry about. [22:25] jderose, I'm working on the Raspberry Pi 2 build. [22:25] flexiondotorg: hey, that's always what i *want* to hear :P [22:25] gotcha [22:26] jderose, Which is a bit unusual because of the kernel. [22:26] Got debug now. [22:26] cyphermox, I'm trying to get oem-config working with my Raspberry Pi 2 build. [22:27] I think the issue is related to that usual kernel I'm using. [22:27] *unusual [22:30] jhodapp: How is it only a partial fix? [22:32] * infinity respins the world. [22:37] althuogh oem-config on lubuntu uses the the unity slideshow but that is probably that there is not a lubuntu oem slideshow package [22:58] ianorlin: no, I think there's just one slideshow for any oem-config. [23:04] There actually is one for ubuntu mate for oem config but apt-cache search slideshow should show them if the package has a reasonable name [23:38] cyphermox, ianorlin There is just one slideshow for owm-config. [23:38] Something I'd like to change for 16.04. [23:39] Or rather, oem-config support only one slideshow. [23:39] ianorlin, Has spotted I tried to create one for Ubuntu MATE in the past. [23:39] cyphermox, I've monkey patched oem-config to work with the Raspberry Pi Foundation kernel. [23:40] cyphermox, No pretty, very specific. I doubt you want a real patch? [23:51] well shoot what you have I can haz a look [23:51] I'm not sleeping just yet.