=== veebers_ is now known as veebers === yofel_ is now known as yofel === _salem is now known as salem_ === salem_ is now known as _salem [05:22] Good morning [07:05] good morning [07:07] hey DanChapman [07:08] hey pitti :-) [07:09] morning DanChapman pitti [07:09] hey elfy [07:47] Good morning [07:52] jibel: I'm currently working on comparing before/after upgrade upstart jobs [07:52] pitti, I'd need latest autopkgtest on stable releases, what do you think is the best option: push the deb to auto-pkg-testing, download it from the archive or publish to backport? [07:52] jibel: http://paste.ubuntu.com/7009176/ is my current diff (I didn't write the post-upgrade test yet) [07:53] jibel: either download and install the deb, or debcheckout autopkgtest and autopkgtest/run-from-checkout [07:53] jibel: we do the latter on the armhf/ppc64el boxes [07:54] jibel: a-u-t> it's a bit unelegant to do the initial recording in bin/auto-upgrade-tester and the post-upgrade recording in UpgradeTestBackendSSH.py, but I didn't see how else it should work [07:54] as in between .upgrade() and .test() the backend is not running, so I can't do the post-upgrade recording in bin/auto-upgrade-tester [07:55] and I can't do it in the post-upgrade test as that doesn't have a reference to the backend [07:55] jibel: does that look acceptable, structurally? [08:07] ugh, wazn is heavily loaded again [08:07] ah, dkms fun [08:07] pitti, why cannot you do the pre-upgrade recording in UpgradeTestBackendSSH.py by adding an upgrade() method and do a super().upgrade() in each backend? would that work too? [08:08] pitti, yeah, it's running the qrt [08:08] pitti, I changed the schedule this morning, there is no need to run it every day [08:08] jibel: ah, I can do that [08:09] there is lot of duplicated code between LXC and Qemu which should be moved to SSH too [08:12] jibel: so more like that? http://paste.ubuntu.com/7009233/ [08:13] pitti, yes, I think it's structurally better [08:13] jibel: added to EC2 as well [08:13] (the super()) [08:14] jibel: thanks [08:14] * pitti will run that on aldebaran, wazn is too busy [08:56] jibel: hm, I see the before/after services in http://d-jenkins.ubuntu-ci:8080/job/upgrade-ubuntu-saucy-trusty-desktop-i386_vm/41/consoleText now, but I think I'll re-structure this again [08:56] jibel: I'll copy the "before" list into the test bed after upgrade, and do the comparison in a post-upgrade test, so that a failure is yellow, not red [08:57] jibel: otherwise I'd have to fail the test in the backend itself, which is ugly [08:57] jibel: or can I raise a test failure (but not an upgrade failure) in the backend, too? [09:18] pitti, it's better to keep before and after in the testbed and run the test there than raising a test failure in the backend [09:18] jibel: *nod* [09:19] jibel: running with that now: http://paste.ubuntu.com/7009420/ [09:21] jibel: meh, now aldebaran has a load of 22; every machine I want to touch today goes crazy :) [09:25] pitti, sounds good. Just a small detail if a service contains 'running' in its name the test might be wrong. [09:26] jibel: right, should check for /running [09:26] or perhaps /running($|,) [09:27] fixed to /running [09:33] meh, I cannot just install autopkgtest from trusty on precise, it depends on debhelper which in turn depends on a whole bunch of packages :( [09:41] jibel: you can't run it from git in your scenario? [09:42] jibel: if you need a package, I can do a precise-compatible backport for you [09:45] pitti, I'd rather use a package that a checkout so we control the version we use [09:45] jibel: the versions are tagged, so you can check out a particular version; but sure, let me build a package [10:01] Morning all [11:00] jibel: I think I'm ready to receive adt spam for ppc64el now; do you want these as well? [11:01] jibel: i. e. I'd revert r298 for ppc64el [11:05] pitti, I do want spam [11:07] jibel: oh, that'll re-enable notifications for everybody, not just us, right? [11:07] jodh: but I guess that's fine [11:08] jibel: committed to r314 [11:09] pitti, or replace {{ email_recipients }} by our emails addresses [11:09] in the templates [11:09] jibel: well, can't hurt if uploaders have a look if they break a test which has worked until now [11:09] sure, they'll feel the pain too [11:14] jibel: http://d-jenkins.ubuntu-ci:8080/job/upgrade-ubuntu-saucy-trusty-desktop-i386_vm/lastSuccessfulBuild/artifact/results/bootstrap.log :) [11:15] services success [11:16] jibel: that's http://paste.ubuntu.com/7009807/ [11:16] jibel: ok to commit and see how it behaves in the other tests? [11:17] pitti, nice. I can't say I succeeded [11:17] + git clone http://anonscm.debian.org/git/autopkgtest/autopkgtest.git /tmp/adt/autopkgtest.checkout [11:17] Cloning into '/tmp/adt/autopkgtest.checkout'... [11:17] error: unable to open object pack directory: /tmp/adt/autopkgtest.checkout/.git/objects/pack: Too many open files [11:17] fatal: failed to read object 198b14448c0dafe33000867fb04bd57d9561fdab: Too many open files [11:17] jibel: ah, I got the same one; weird with a proxy [11:17] jibel: a git pull after you have a checkout generally succeeds, but I've seen this on the initial clonse [11:18] jibel: workaround: clone locally, scp -r [11:26] more and more interesting fatal: unable to access 'http://anonscm.debian.org/git/autopkgtest/autopkgtest.git/': Failed to connect to anonscm.debian.org port 80: Connection refused [12:01] balloons: https://code.launchpad.net/~elfy/ubuntu-manual-tests/1229511/+merge/208780 [12:04] balloons, was your update-tracker-from-branch-script public and something others could use as well? === _salem is now known as salem_ [13:23] rhuddie: you can put your MP as ready for review. [13:24] elopio, will do [13:28] elopio, I can't find the option for that, do I need to propose for merge again? [13:28] rhuddie: no, you only need to propose the merge again when you change prerequirements. [13:29] rhuddie: https://code.launchpad.net/~rhuddie/address-book-app/test_go_to_add_contact/+merge/206828 [13:29] on top, it says status: Work in progress [13:29] you should have a yellow button next to it. [13:30] elopio, ah thanks. easy when you know how :) [13:30] rhuddie: that's launchpad's motto. === salem_ is now known as _salem === _salem is now known as salem_ === alex_abreu is now known as alex-abreu [15:11] rvr, jcollado, rhuddie: want to hangout today? [15:12] elopio, sounds good to me [15:12] elopio: Yeah [15:13] elopio: Ok [15:13] teenagers... ;) [15:13] rhuddie, rvr, jcollado: ok, same place https://plus.google.com/hangouts/_/canonical.com/new-features [15:13] going to hang out in the mall? [15:13] * knome hides [15:14] and random time, as always :) [15:14] (for the record, i was just kidding) [15:30] knome: the cool kids hangout without getting up from the couch. [15:31] you mean the fat kids? [15:31] * knome hides [15:31] ow, i can't fit under the table! [15:31] ;) [15:44] lol [16:06] * elopio goes to make some excercise :'( [16:06] elopio: These are the failures in webapps-demo on the phone with a clean install http://paste.ubuntu.com/7011100/ [16:06] Are they similar to yours? [16:07] rvr: yes, that's what I'm getting. [16:07] "returned multiple results" [16:07] I will take a look [16:08] That's supposed to be fixed by veebers [16:08] rvr: on his branch, alex changed the way we launch the apps: [16:08] 7 + self.app_proxy = self.launch_click_package( [16:08] 218 + package_name, [16:08] 219 + app_name) [16:08] DanChapman, balloons, hi :) [16:08] that might do the trick. [16:09] thanks for looking into it rvr. Let me know if you find something. === salem_ is now known as _salem === _salem is now known as salem_ === bfiller is now known as bfiller_afk === bfiller_afk is now known as bfiller [21:24] balloons, elopio hi === salem_ is now known as _salem