=== salem_ is now known as _salem
=== chihchun_afk is now known as chihchun
[08:07] <pitti> jibel: did you try prepare-testbed locally recently? since yesterday I get a VM without autopkgtest, dpkg-dev etc. installed; but the command succeeds and the log says it did install those
[08:07] <pitti> would be interesting to see if that's my system or cloud-init
[08:08] <jibel> pitti, I didn't recently, I'll have a look
[08:36] <jibel> pitti, that's interesting, my proxy shows that additional packages have been fetched but they are not installed in the VM
[08:41] <pitti> jibel: yeah, seems to be the same here
[08:42] <jibel> there is no recent change to cloud-init, apt log is empty, and no dpkg log
[08:44] <jibel> pitti, I'm disabling the script that updates VM for autopkgtest
[08:44] <jibel> until we figure what's wrong
[08:44] <pitti> jibel: you mean re-create?
[08:44] <jibel> yes
[08:44] <pitti> the dist-upgrade ought to work
[08:44] <pitti> ack
[08:44] <pitti> jibel: although I'm surprised that it didn't affect the VMs in the DC
[08:45] <pitti> jibel: at first I thought that's a problem on my side as I didn't see tests failing due to that ("adt-run: not found"
[08:45] <pitti> I guess they were recreated this morning, as usual
[08:46] <pitti> ok, in the DC I can trivially reproduce the espeak-data bug
[08:48] <jibel> pitti, you're right, VMs have been recreated in the lab this morning, and they don't have this problem. So that'd be something on the host's side.
[08:49] <pitti> no /var/log/dpkg.log at all here
[08:50] <pitti> but cloud-init.log does have the apt-get dpkg-dev & friends
[08:55] <jibel> I don't, in my case it seems to abort in the middle of apt-upgrade and jumps to boot-finished
[08:58] <jibel> VM of saucy are prepared correctly too
[09:04] <jibel> pitti, ah /dev/vda1        7191032 7174648         0 100% /
[09:04] <jibel> bash: cannot create temp file for here-document: No space left on device
[09:04] <jibel> I saw a recent change to grow_part in trusty
[09:04] <pitti> ah, hah!
[09:04] <jibel> that's probably that
[09:06] <pitti> so the Tahr ought to be put on a diet? :-)
[09:07] <pitti> Ubuntu 14.04 LTS "Chubby Tahr"
[09:12] <jibel> pitti, I miss something there http://paste.ubuntu.com/6790575/
[09:12] <jibel> according to du there is only 1G in use
[09:13] <pitti> some dot files?
[09:13] <pitti> jibel: where do you run that, in run-adt-test -sUl?
[09:13] <jibel> pitti, during the provisioning
[09:14] <jibel> I added a sleep 3600 to wait_bootfinished and ssh-ed into the VM
[09:17] <pitti> doing the same now
[09:20] <pitti> $ sudo du -hsx /
[09:20] <pitti> 817M/
[09:21] <pitti> hmm
[09:23] <jibel> pitti, if a file is deleted but a process still writing to it, would it be visible in lsof?
[09:24] <pitti> it should
[09:24] <pitti> jibel: at least you see it in /proc/pid/fd, there's a symlink that says -> /destination/path (deleted)
[09:25] <pitti> $ sudo ls -l /proc/*/fd/ | grep deleted
[09:25] <pitti> $
[09:25] <pitti> jibel: maybe the journal went out of its mind and grew way too much?
[09:26] <pitti> Journal size:             32M
[09:26] <pitti> hm, seems ok
[09:27] <pitti> Block count:              1572608
[09:27] <pitti> Free blocks:              1340686
[09:27] <pitti> Free inodes:              332101
[09:28] <pitti> jibel: I think this coincides fairly well with 3.13.0-4
[09:32] <pitti> jibel: FTR; you see deleted files in lsof; I started a flash video locally, and both /proc/<pid>/fd/ and lsof show the deleted files
[09:44] <pitti> jibel: hm, we don't seem to have a cloud image with 3.13.0-3 any more :/
[09:45] <jibel> pitti, and I think we keep an history of server and desktop images only
[09:45] <pitti> jibel: I meant, even http://cloud-images.ubuntu.com/trusty/20140118/ already has -4
[09:46] <jibel> pitti, right, and I mean in the QA lab we keep an history of daily images for several months but only for server and desktop
[10:19] <slickymaster> morning all
=== _salem is now known as salem_
[10:55] <pitti> http://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-urdfdom/1/ARCH=i386,label=adt/console
[10:55] <pitti> "/home/ubuntu/run-adt: 255: /home/ubuntu/run-adt: Syntax error: redirection unexpected"
[10:55] <pitti> jibel: why is this hating us so much?
[10:55]  * pitti queues that after espeak-data
[10:56] <pitti> line 255 just says "fi", WTF
[10:56] <pitti> there is no redicrection in the vicinity
[10:56] <jibel> ???
[10:57] <jibel> is it confused by the >= in expr?
[10:57] <pitti> hardly, it's quoted
[10:57] <pitti> first time I saw it
[10:58] <pitti> I just pretend I didn't see it, I retried it for now
[10:58] <davmor2> Morning all
[11:07] <jibel> pitti, I'm working on this run-adt failure. It is reproducible
[11:07] <pitti> oh, is it?
=== chihchun is now known as chihchun_afk
[11:10] <jibel> pitti, er, it is caused by a conflict
[11:10] <jibel> <<<<<<< TREE
[11:10] <jibel> :/
[11:10] <jibel> fixing
[11:10] <pitti> ah, haha
[11:11] <pitti> jibel: so it's because wazn takes the brunt of the tests, that errors on alderamin/aldebaran only show up occasionally
[11:11] <jibel> pitti, actually only on albali
[11:12] <pitti> hm, it seems my interweb tube to wazn froze a few minutes ago
[11:12] <jibel> resolved
[11:12] <pitti> jibel: merci
[11:15] <pitti> jibel: can't reach wazn ATM; I'll retry and clean up my ephemeral container once I get back there
[11:16] <jibel> pitti, I notified the CI team
[11:16] <jibel> psivaa is on it
[11:16] <pitti> oh, it's actually broken?
[11:16] <pitti> thanks
[11:39] <pitti> jibel: do you know which kernel/release wazn runs?
[11:39] <pitti> jibel: I tried to build a local quantal container for bug 1220681, but I can't reproduce it locally under trusty
[11:40] <ubot5> bug 1220681 in espeak (Ubuntu) "package espeak-data 1.46.02-2ubuntu1 failed to install/upgrade: unable to move aside `./usr/lib/i386-linux-gnu/espeak-data/voices/en' to install new version: Invalid cross-device link" [Undecided,Confirmed] https://launchpad.net/bugs/1220681
[11:41] <pitti> jibel: it's raring, isn't it?
[11:44] <pitti> err, saucy
=== chihchun_afk is now known as chihchun
[11:52] <pitti> jibel: jenkins is confused now
[11:53] <pitti> http://d-jenkins.ubuntu-ci:8080/job/trusty-adt-gtk-3.0/58/ is FAIL although both arches succeeded, and same with http://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-urdfdom/
[11:57] <jibel> meh FATAL: hudson.remoting.RequestAbortedException: java.net.SocketException: Socket closed
[11:58] <jibel> it won't confuse britney though because it collects individual results
[11:58] <pitti> ah, good
[12:02] <pitti> jibel: so today, the kernel, autopkgtest, jenkins, and overlayfs cooperate to screw us
[12:02] <pitti> Heghlu'meH QaQ jajvam
[12:34] <jibel> pitti, wazn is back
[12:35] <jibel> it runs saucy with kernel  3.11.0-15-generic #23-Ubuntu
[13:26] <pitti> jibel: thanks; I tried that kernel, still can't reproduce localy
[13:55] <pitti> jibel: did you already happen to file a bug against unity for the python:any dependency upgrade failure (http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-precise-trusty-desktop-i386_vm/2/console) ?
[13:55] <pitti> jibel: if not, I'll have a look
[13:56] <jibel> pitti, no, not yet, I'm on X not starting in LXc after the upgrade
[13:56] <pitti> jibel: ack, taking that then
[13:56] <pitti> looks like a missing pre-dependency on python or dpkg, I'll check with doko
=== chihchun is now known as chihchun_afk
[14:19] <pitti> jibel: espeak-data workaround uploaded FYI
[14:19] <pitti> (should fix quantal -> trusty upgrades)
=== chihchun_afk is now known as chihchun
[15:16] <balloons> davmor2, more requests for you to run things on maguro :-) Can you give this branch a whirl? lp:~nskaggs/ubuntu-clock-app/fix-testtimer-tests
[15:16] <balloons> you can try out sergio's click buddy tool if you've not seen it https://code.launchpad.net/~sergiusens/phablet-tools/click-buddy/+merge/202178
[15:17] <davmor2> balloons: currently bisecting an issue on the maguro so it might take some time
[15:17] <balloons> ahh.. no worries
=== salem_ is now known as _salem
=== spineau is now known as spineau_afk
[16:59] <elfy> hi balloons - I see your comment on that 'English' bug - I've made a start on those
[17:00] <balloons> hey elfy.. Excellent. I trust your wordsmithing!
[17:00] <elfy> you sure boss?
[17:00] <elfy> :p
[17:03] <balloons> lol.. ofc ofc
[17:04] <elfy> I'm starting with UK English - everything else pales into insignificance, knock yourself out
=== Ursinha is now known as Ursinha-afk
=== rbasak_ is now known as rbasak
=== Ursinha-afk is now known as Ursinha
=== _salem is now known as salem_
[18:40] <elopio> balloons: this is what I've been able to do:
[18:40] <elopio> $ account-console create --print-id ubuntuone
[18:40] <elopio> 18
[18:40] <elopio> $ account-console edit --username u1test+20140121@canonical.com --password Hola123* 18
[18:40] <elopio> OK 18
[18:40] <elopio> $ account-console login 18
[18:40] <elopio> {}
[18:40] <elopio> Got reply: {'UserName': 'u1test+20140121@canonical.com', 'Secret': 'Hola123*'}
[18:40] <elopio> I'm still missing something.
[18:42] <balloons> elopio, I was chatting with ken and mardy about it. There's no way around going through the browser prompt apparently
[18:42] <elopio> balloons: I get no browser prompt.
[18:43] <balloons> account-console login 18
[18:43] <elopio> I suppose you are one step ahead.
[18:45] <balloons> lol.. so the suggestion by cwayne was to automate going through settings and adding an account
[18:45] <balloons> if we make it generic enough it could be re-used
[18:47] <elopio> balloons: That's what I'm trying. I'm looking at the status of their helpers.
[18:47] <elopio> I worked on the ones to add an u1 account, so that should be usable.
[18:48] <balloons> where at?
[19:00] <elopio> balloons: http://bazaar.launchpad.net/~ubuntuone-control-tower/ubuntuone-credentials/trunk/view/head:/online-accounts-provider/tests/autopilot/UbuntuOneCredentialsProviderAutopilotTests/emulators.py
[19:00] <elopio> hum, I guess we need a deb for it.
[19:05] <balloons> hmm..
=== chihchun is now known as chihchun_afk
=== salem_ is now known as _salem