[04:51:46] https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#WMF_donation_appeal_blocked_my_viewing_of_Wikipedia.2C_with_no_option_to_dismiss [04:54:05] thedj, is that the page that clicking on the banner takes you to? [04:54:29] * thedj has no clue.... [04:55:27] I doubt that would be a deliberate action from FR [12:28:05] wm-labs-meetbot: [12:56:19] CI weekly meeting about to start here . Feel free to join https://lists.wikimedia.org/pipermail/qa/2015-March/002213.html [13:02:06] o/ [13:02:07] #startmeeting CI weekly meeting [13:02:07] Meeting started Mon Mar 30 13:02:07 2015 UTC and is due to finish in 60 minutes. The chair is hashar. Information about MeetBot at http://wiki.debian.org/MeetBot. [13:02:07] Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. [13:02:07] The meeting name has been set to 'ci_weekly_meeting' [13:02:11] o/ [13:02:17] zeljkof: attending ? [13:02:23] hashar: lurking :) [13:02:24] jzerebecki: you can raise your hand again :D [13:02:26] ok [13:02:32] \o [13:02:36] going to be a lonely meeting I guess [13:02:41] #link Agenda https://www.mediawiki.org/wiki/Continuous_integration/meetings/2015-03-30 [13:02:51] #topic actions retrospective [13:02:52] #info No previous meeting [13:03:09] #topic next meeting [13:03:09] Next Monday is easter. Reschedule to Tuesday April 7th same time? [13:03:19] #action Antoine to update calendar invite for next week [13:03:43] jzerebecki: we can actually add the disk space consumption as a topic [13:03:54] #topic disk space exhaustion on labs slaves [13:04:05] jzerebecki just noticed integration-slave1001 got a full /mnt/ [13:04:37] following the migration of jobs to unified jobs (such as 'npm'), the old workspaces have not been collected [13:05:01] we would need an easy way to detect obsolete workspaces on all slaves and collect them [13:05:12] #action Antoine to fill a task regarding workspaces garbage collection [13:06:14] A bunch of jobs have a full clone of mediawiki/core which is a full copy. That takes unnecessary disk space [13:06:18] #link https://phabricator.wikimedia.org/T93703 [13:07:05] we can be smarter by having the workspace clone from a local mirror of mw/core. With git clone --shared, that would point .git/objects/info/alternates to that local mirror and save a huge amount of disk on our slave [13:07:37] #action Need a champion to review/test/implement/deploy the git clone --shared described at https://phabricator.wikimedia.org/T93703#1144542 [13:07:58] * hashar feels lonely :D [13:08:16] #topic CI isolation status [13:08:31] #link https://phabricator.wikimedia.org/project/board/1010/ [13:09:00] Zuul Debian package ready for Precise and Trusty. Almost done, pending final approval by ops [13:09:17] Antoine feels it is ready for deployment on one of our labs instance. [13:09:38] #action Antoine to deploy zuul package on a labs instance and generalize on all instances [13:10:43] #info Antoine giving a short presentation of the project to ops on Wednesday April 1st [13:11:44] #info Project could be fulfilled with a single new server instead of two. See procurement ticket https://phabricator.wikimedia.org/T93076 [13:11:53] #topic Publish demo dir to doc.wikimedia.org [13:12:33] #link https://phabricator.wikimedia.org/T62143 [13:12:54] VE has a /demos/ standalone directory that they would like to see published under doc.wikimedia.org [13:13:11] it is fully client-side [13:14:43] #info We should be able to craft a job that triggers doc-publish-sync job with source /demos/ and dest /visualeditor//demos/ [13:15:21] #info Not much time on Antoine side right now though. That task needs a volunteer. [13:15:28] #topic Publish QUnit coverage [13:15:33] #link https://phabricator.wikimedia.org/T87490 [13:16:16] seems to be [13:16:38] seems to be grunt work like previous topic. Have to create a job that triggers doc-publish-sync with appropriate source/dest parameters [13:16:47] ooh nice [13:17:35] #info same as previous topic: need a volunteer to craft a post merge job that will generate the coverage and publish from /dist/coverage/ to /cover/{project} [13:17:45] jzerebecki: yeah there are some repos using coverage [13:17:59] but publishing them on doc.wikimedia.org is not as easy as it should :( [13:18:25] we have a job in JJB that ease the configuration (defined in integration/config.git /jjb/doc.yaml [13:19:07] the process is roughly: generate stuff, trigger doc-publish-sync job with the dir to publish in the workspace and the destination under doc.wikimedia.org [13:19:36] #link https://git.wikimedia.org/blob/integration%2Fconfig/master/jjb%2Fdoc.yaml [13:19:54] #topic l10n-bot self merging [13:19:58] #link https://phabricator.wikimedia.org/T91707 [13:21:08] what the hell is happening [13:21:31] the l10-bot has been moved to use only CR+2 [13:21:43] but reverted bah [13:22:35] #info l10n-bot switched last week to use CR+2 but bunch of repositories are failing tests. Current situation is no more tests are run at all. [13:23:21] #action Figure out what we really want to do when l10n bot propose a patch. Antoine in favor of skipping tests on patch upload then CR+2 and gate-and-submit to merge patches. [13:23:32] #topic Gerrit change merged message [13:24:08] #link https://phabricator.wikimedia.org/T91766 [13:26:22] #info That is related to Gerritbot updating the Phabricator task whenever a patch is merged in. Not really linked to CI [13:26:43] #action https://phabricator.wikimedia.org/T91766 is for Gerritbot, not directly CI so leave it in #wikimedia-git-or-gerrit [13:26:52] #topic Automatic config for extensions [13:26:57] #link https://phabricator.wikimedia.org/T92909 [13:28:15] #info Extensions are created in Gerrit but almost never have an initial set of CI jobs set for them. The task propose to have a wildcard to run a default set of jobs, that is made possible with the overhauling of job names which are almost indentical between repositories. [13:28:58] #info We will need add support for wildcard in Zuul layout. Need to be able to update/patch Zuul which depends on having it deployed via Debian package [13:29:13] #action Antoine to make the task blocked by Zuul package deployment [13:29:30] #info Need a volunteer to write the python code and upstream it to OpenStack [13:29:50] #topic Triage [13:29:55] #link https://phabricator.wikimedia.org/project/board/401/query/open/ [13:30:23] #info No point doing a triage alone :) [13:30:39] #topic Wikidata [13:30:40] jzerebecki: do you want to add some #info regarding wikidata? [13:30:50] since you guys apparently have been rather active recently [13:31:12] mh [13:31:22] no need to write a novel [13:31:32] if you have in mind a few items you can #info them :) [13:31:46] so whatever progress you made will be public / archived etc [13:31:47] was not really prepared for this [13:31:57] I can imagine [13:32:07] will do better next week :] [13:33:14] #info WikidataTests browser test job now timeout after 4 hours. [13:33:40] we were having problems of stuff being blocked on review by someone with +2 on ci stuff, but i think all of those were commented on and those were addressed already... so it moves :) [13:33:53] #info The ridiculous long run time is due to our test runner taking 20-40 seconds before each scenario being run. See https://phabricator.wikimedia.org/T92613 [13:34:03] great! [13:34:23] jzerebecki: hopefully this weekly meeting will give you an occasion to raise blocked tasks [13:34:34] so they can get reviewed [13:34:38] yes that is a good idea [13:35:18] #info Some changes been blocked lacking review from +2ers. Apparently all sorted out. This meeting will be a good way to raise awareness on such blocking tasks. [13:35:38] jzerebecki: I haven't been very active for the last few weeks and timo / kunal been busy overhauling the job names [13:35:44] so that explain the review lag [13:35:54] I think I am done dumping my brain [13:36:24] hashar: can you invite jan.zerebecki@gmail.com to the calendar item for this? [13:36:32] #topic misc [13:37:01] #info Next meeting Tuesday April 7th (Monday is Easter) [13:37:53] jzerebecki: I think I have invited you :) [13:38:02] thank you to have lurked [13:38:10] next week will get more people hopefully! [13:38:43] hashar: yes thx [13:38:46] I will post the minutes on the wiki at https://www.mediawiki.org/wiki/Continuous_integration/meetings/2015-03-30 [13:39:06] jzerebecki: we can talk about browser test in this meeting as well [13:39:10] if needed [13:39:23] #endmeeting [13:39:23] Meeting ended Mon Mar 30 13:39:22 2015 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) [13:39:23] Minutes: https://tools.wmflabs.org/meetbot/wikimedia-office/2015/wikimedia-office.2015-03-30-13.02.html [13:39:23] Minutes (text): https://tools.wmflabs.org/meetbot/wikimedia-office/2015/wikimedia-office.2015-03-30-13.02.txt [13:39:23] Minutes (wiki): https://tools.wmflabs.org/meetbot/wikimedia-office/2015/wikimedia-office.2015-03-30-13.02.wiki [13:39:23] Log: https://tools.wmflabs.org/meetbot/wikimedia-office/2015/wikimedia-office.2015-03-30-13.02.log.html [13:39:31] lovely meetbot [13:39:57] zeljkof: done :) [13:40:03] zeljkof: next week will be more active I guess [13:40:13] hashar: just the two of you active? [13:42:59] yeah [13:43:06] zeljkof: the time is not ideal [13:43:07] hashar: that's a start :) [13:43:15] and I announced it only on Friday [13:43:24] I learned a few things though [13:43:34] it is good to have all the agenda prepared ahead of time [13:43:39] and all the meetbot command ready [13:43:46] https://tools.wmflabs.org/meetbot/wikimedia-office/2015/wikimedia-office.2015-03-30-13.02.html !!! [15:02:32] * Coren idly wonders if YuviPanda has strange urges to wear 10 gallon hats and bear arms now that he's a 'murican. :-) [15:02:56] (Well, adopted temporary 'murican) [15:34:48] Coren: lol :) [15:39:11] Coren: Well, right now he's living in the house of two Brits, so in fact his experience of the US so far has been people with a Northern English accent trying to convince him to eat baked beans. [15:39:46] Hah! [15:47:51] (He wouldn't, by the way) [15:48:04] I said all that time he's spent in the UK recently has gone to waste if he's still not eating beans.