| === czchen_ is now known as czchen | ||
| === TheLordOfTime is now known as teward | ||
| pitti | Good morning | 05:37 |
|---|---|---|
| pitti | jibel: ok to push http://paste.ubuntu.com/6913682/ to lp:auto-upgrade-testing? | 07:57 |
| jibel | pitti, ok | 07:57 |
| pitti | jibel: I was running the script on my own box to see whether there's anything wrong with it, but it's fine aside from that | 07:58 |
| pitti | jibel: done | 07:58 |
| pitti | jibel: so it seems that that upgrades don't actually run due to that authentication error, and the only thing that tells us is that kernel test | 07:59 |
| jibel | pitti, right, and only in the lab. I'm pretty sure keys or tarball are cached somewhere on the network but couldn't figure exactly where. | 08:01 |
| pitti | jibel: that's on wazn, right? I'm having a look | 08:01 |
| jibel | pitti, and retoaded told me there is no proxy and squid is configured to not cache these files | 08:01 |
| jibel | pitti, yes on wazn | 08:01 |
| pitti | so, I did set $http_proxy on wazn recently, but upgrades failed before that with the same error | 08:02 |
| jibel | pitti, same error happens for tests running in VMs which are running on alderamin | 08:02 |
| jibel | pitti, it is unrelated to your change | 08:03 |
| pitti | jibel: right, the containers don't have $http_proxy | 08:08 |
| === maclin_ is now known as maclin | ||
| pitti | jibel: meh: trying that in a local schroot: http://paste.ubuntu.com/6913730/ | 08:10 |
| pitti | ah, python not installed, and there's no dependency on it | 08:12 |
| jibel | pitti, yes you need python-apt | 08:14 |
| pitti | so, in my schroot it's working fine | 08:15 |
| pitti | err, WTF? | 08:19 |
| pitti | jibel: so yes, trusty.tar.gz is from Oct 29, trusty.tar.gz.gpg from Jan 24 | 08:25 |
| pitti | wgetting http://archive.ubuntu.com/ubuntu/dists/trusty/main/dist-upgrader-all/current/trusty.tar.gz in the container works fine | 08:29 |
| * pitti tracks down how that tarball gets downloaded in release-upgrader | 08:29 | |
| jibel | pitti, it is clearly a problem with the proxy | 08:31 |
| jibel | pitti, if you remove Acquire::http:proxy from /etc/apt.conf it should work | 08:31 |
| pitti | $ sudo apt-config dump|grep -i proxy | 08:32 |
| pitti | Acquire::http::proxy "http://10.98.3.6:8000/"; | 08:32 |
| pitti | ah, yes | 08:32 |
| jibel | I'm pretty sure squid-deb-proxy is misconfigured | 08:32 |
| jibel | and caches tar files and/or gpg keys | 08:32 |
| jibel | but I cannot check | 08:32 |
| pitti | $ http_proxy=http://10.98.3.6:8000/ wget http://archive.ubuntu.com/ubuntu/dists/trusty/main/dist-upgrader-all/current/trusty.tar.gz | 08:33 |
| pitti | 1125177 Oct 29 18:28 trusty.tar.gz | 08:33 |
| pitti | yep | 08:33 |
| jibel | I'll ask CI again | 08:33 |
| pitti | should I send a ticket? | 08:33 |
| jibel | pitti, yes please, but Larry told me it ignore these files which it obviously doesn't | 08:34 |
| jibel | and we cannot upgrade without a proxy, it would take ages only to download packages | 08:35 |
| pitti | jibel: sent | 08:38 |
| pitti | jibel: I'd look for a workaround which temporarily disables the proxy for downloading the release tarball | 08:38 |
| jibel | pitti, thnaks, I asked cihelp to clear the cache too | 08:38 |
| jibel | pitti, but you need to patch u-r-u for that? | 08:39 |
| pitti | jibel: well, we could locally install apt-cacher-ng, but if it can be fixed properly on tachas that'd be better of course | 08:39 |
| pitti | jibel: I don't want to do that, of course; I'll think about it | 08:40 |
| pitti | jibel: hm, right, you can't really wedge a configuration change in between downloading the tarball and the debs, darn | 08:44 |
| pitti | eek, why are so many adt jobs hanging for a long time, /me investigates | 08:45 |
| pitti | conffile prompt | 08:45 |
| pitti | jibel: ^ filed bug 1278770 FTR | 08:49 |
| ubot5 | bug 1278770 in pollen (Ubuntu) "upgrading pollinate causes conffile prompt" [High,New] https://launchpad.net/bugs/1278770 | 08:49 |
| * pitti cleans up after it | 08:49 | |
| pitti | jibel: can you please bzr pull a-pkg-testing on albali? I committed a workaround for this ^ (--force-confnew) | 09:23 |
| pitti | jibel: sorry that I keep annoying you with this stuff :( | 09:24 |
| jibel | pitti, no problem. pulled r280 | 09:25 |
| pitti | cheers | 09:25 |
| jibel | pitti, you might create a jenkins job called "shell" that takes a shell command in argument :) | 09:25 |
| pitti | heh | 09:25 |
| davmor2 | Morning all | 10:55 |
| pitti | jibel: can you please roll out a-pkg-testing r281 on albali? I think I fixed the problem now | 11:23 |
| pitti | jibel: prepare-testbed failed on albali (as pitti) and now works | 11:23 |
| jibel | pitti, done | 11:24 |
| pitti | jibel: cheers; reenabling albali then | 11:25 |
| pitti | argh, wazn still offline | 11:26 |
| === _salem is now known as salem_ | ||
| pitti | jibel: yay! http://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-setup-testbed/109/? | 11:51 |
| jibel | pitti, woohoo. What was it, slow to stop or not stopped at all? | 11:52 |
| pitti | jibel: apparently just slow to stop; and I never noticed on the three others or on my machine | 11:52 |
| pitti | jibel: I now added a proper wait loop for the process to go away | 11:52 |
| pitti | jibel: now of course wazn failed :) | 11:53 |
| pitti | infrastructure hates me today | 11:53 |
| jibel | pitti, k, the disk setup is different on albali and they are very slow compared to others | 11:53 |
| jibel | pitti, not only you :) | 11:53 |
| pitti | that's music in my ears: http://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-setup-testbed/109/ | 12:56 |
| pitti | jibel: ^ :) | 12:56 |
| jibel | pitti, das ist wunderbar! | 12:58 |
| pitti | jibel: now test_old_kernel_is_installed fails because the dist-upgrader actually cleans up the old kernel | 13:41 |
| pitti | jibel: I'm not entirely sure whether that is intended | 13:42 |
| jibel | pitti, I noticed that too, since we are running in LXC I am not sure if this behavior is correct or not | 13:42 |
| pitti | jibel: ah, I see | 13:43 |
| pitti | jibel: it would see that 3.11 is not the running kernel, so apt's autoclean would remove it | 13:43 |
| pitti | jibel: let me try the -VM one | 13:43 |
| jibel | I suppose so | 13:43 |
| pitti | jibel: xorg-is-running fails, too | 13:43 |
| pitti | self.assertTrue(is_process_running('Xorg'), 'X Server not running') | 13:43 |
| pitti | that's wrong even on my trusty box | 13:43 |
| pitti | jibel: should probably be 'X', not 'Xorg'? | 13:44 |
| === alex__ is now known as alex-abreu | ||
| jibel | pitti, yes, has it ever been Xorg? | 13:44 |
| pitti | jibel: not recently at least | 13:44 |
| pitti | jibel: but to be sure we could do is_process_running('Xorg') or is_process_running('X') | 13:45 |
| jibel | maybe in Precise, I cnanot remember | 13:45 |
| * pitti boots current precise live system | 13:46 | |
| pitti | hm, but I have a 12.04.4 version with LTS enablement, that's already the saucy stack | 13:46 |
| pitti | (it's X, too) | 13:46 |
| pitti | jibel: can I push the change to just "X" and we see how it goes? | 13:46 |
| jibel | pitti, of course | 13:47 |
| jibel | and there is nothing to update on albali :) | 13:47 |
| pitti | lol | 13:47 |
| pitti | err, what? | 13:47 |
| pitti | $ share/post_upgrade_tests/test_xserver.py -v | 13:47 |
| pitti | Checking for running Xorg ... skipped 'Not a desktop installation.' | 13:47 |
| pitti | oh whoops, I indeed don't have ubuntu-desktop installed | 13:48 |
| pitti | hmm, the test succeeds here, /me investigates | 13:49 |
| pitti | $ ps -eo comm|grep X | 13:49 |
| pitti | Xorg | 13:49 |
| pitti | jibel: so presumably because we start it in a container? | 13:49 |
| jibel | pitti, the process should exist anyway, even with the dummy driver | 13:50 |
| pitti | jibel: WDYT: paste.ubuntu.com/6915122/ | 13:55 |
| pitti | also observe the != → >= (we don't want to pass on zero kernels ☺ ) | 13:55 |
| jibel | pitti, LGTM | 14:04 |
| pitti | jibel: done and rolled out | 14:10 |
| === om26er_ is now known as om26er | ||
| bfiller | jibel: hello | 14:20 |
| jibel | bfiller, Hey | 14:20 |
| bfiller | jibel: can you update your MR for address-book, getting a merge conflict: http://162.213.34.102/job/landing-002-1-build/23/console | 14:21 |
| jibel | bfiller, yes, it's on my list for today. | 14:22 |
| bfiller | jibel: ok thanks, trying to get that through the "train" | 14:24 |
| jibel | bfiller, good luck with that. An update to the existing branch would be fine or you need a new MP? | 14:25 |
| bfiller | jibel: just update to existing branch, think you just need to bzr merge first with trunk and resolve the conflict and then repush | 14:25 |
| pitti | jibel: http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-saucy-trusty-desktop-amd64_vm/21/artifact/results/bootstrap.log looks fairly good now | 14:27 |
| pitti | jibel: coudl it be that we don't reboot the VM after upgrade before we run the post-upgrade tests? | 14:27 |
| jibel | pitti, we are rebooting, if it doesn't it's a bug | 14:30 |
| pitti | jibel: ack, thanks | 14:30 |
| jibel | pitti, "QEMU 1.0 monitor - type 'help' for more information" near the ends suggests it did | 14:30 |
| jibel | -s | 14:30 |
| pitti | right, just saw that | 14:30 |
| pitti | jibel: hm, it looks like 3.13 got installed and grub was updated accordingly; I'll reproduce that on wazn manually and check | 14:34 |
| pitti | (also the X.org in LXC thing), later this afternoon | 14:34 |
| jibel | bfiller, actually it is not a conflict with trunk but 2 MPs | 14:34 |
| bfiller | jibel: two conflicting MP's that are trying to be merged? | 14:46 |
| jibel | bfiller, yes, not sure how to fix that. Could you merge https://code.launchpad.net/~jibel/address-book-app/buttons_moved_to_emulators first, then I'll rebase the second MP on the new trunk? | 14:47 |
| jibel | I cannot stack them because they're unrelated | 14:48 |
| bfiller | jibel: let me try | 14:52 |
| senan | Good Evening :) | 16:28 |
| senan | DanChapman,balloons ; Hey :) | 16:29 |
| balloons | senan, hello | 16:29 |
| balloons | DanChapman, evening to you as well. I actually have some news for you | 16:29 |
| balloons | DanChapman, http://www.tech-foo.net/delayed-assertions-in-python-testtools.html | 16:29 |
| pitti | jibel: where on wazn do I actually find the upgrade tsting VMs? | 16:34 |
| pitti | jibel: there are two images in back/ but they are alreayd half a year old and are named saucy*.raring, so presumably they are old | 16:34 |
| pitti | jibel: ooh, they are on aldebaran, not on wazn | 16:38 |
| pitti | jibel: so, what's the difference between test-image.ubuntu-saucy-trusty-desktop-i386{,.saucy}? | 16:39 |
| jibel | pitti, VMs are on alderamin | 16:41 |
| pitti | oh, http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-saucy-trusty-desktop-amd64_vm/21/consoleFull says "aldebaran-upgrade" | 16:41 |
| jibel | let me check | 16:42 |
| jibel | ah right, precise is on alderamin and saucy on aldebaran | 16:43 |
| jibel | it'd be easier if the server let me in | 16:43 |
| jibel | ... | 16:43 |
| pitti | jibel: ah, I think .saucy is the original, and the other the upgraded VM | 16:45 |
| jibel | pitti, so .saucy is the original | 16:45 |
| jibel | adn the other is the result | 16:45 |
| jibel | what you said :) | 16:45 |
| pitti | thanks | 16:45 |
| pitti | $ uname -a | 16:46 |
| pitti | Linux ubuntu 3.11.0-15 | 16:46 |
| pitti | en effet | 16:46 |
| pitti | vmlinuz -> boot/vmlinuz-3.13.0-8-generic | 16:47 |
| pitti | in /boot/grub/menu.list there is indeed just 3.11 | 16:47 |
| pitti | /boot/grub/grub.cfg has 3.13 | 16:48 |
| pitti | looks like the boot sector is still using the old grub perhaps? | 16:48 |
| jibel | pitti, is there only grub2 install or there is another older version? | 16:51 |
| jibel | +ed | 16:51 |
| senan | DanChapman, can you please check this http://paste.ubuntu.com/6915973/ | 16:53 |
| senan | :) | 16:53 |
| pitti | jibel: re, checking | 17:02 |
| pitti | jibel: only grub 2 | 17:02 |
| pitti | jibel: hm, I ran install-grub /dev/hda and rebooted, I think I ruined the VM (doesn't come up) | 17:04 |
| pitti | but, that'll just be re-created at the next dist-upgrade run anyway, right? | 17:04 |
| jibel | pitti, yes, it is just kept after the test to troubleshoot upgrade failures | 17:05 |
| * pitti re-runs amd64 in jenkins and logs into the i386 VM instead | 17:07 | |
| pitti | ah, that's still saucy, due to the tachash bug; /me rebuilds, too | 17:09 |
| pitti | jibel: so the idea is that the containers have lightdm running, which should start X with the dummy driver? | 17:27 |
| jibel | pitti, that's the idea | 17:30 |
| pitti | jibel: hm, so it's not even running in the un-upgraded container | 17:31 |
| pitti | lightdm is, but not X | 17:31 |
| pitti | (EE) Server terminated successfully (0). Closing log file. | 17:31 |
| pitti | how helpful | 17:31 |
| pitti | [+0.50s] DEBUG: Session pid=941: Started with service 'lightdm', username 'ubuntu' | 17:32 |
| pitti | [+0.50s] DEBUG: Got signal 15 from process 766 | 17:32 |
| pitti | [+0.50s] DEBUG: Caught Terminated signal, shutting down | 17:32 |
| pitti | [+0.50s] DEBUG: Stopping display manager | 17:32 |
| senan | DanChapman, Can you please check the ringchart and treemap issue ? | 17:38 |
| === bfiller is now known as bfiller_afk | ||
| === bfiller_afk is now known as bfiller | ||
| === salem_ is now known as _salem | ||
| === charles_ is now known as charles | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!