[00:40:05] Are "experimental" jobs in zuul/layout.yaml protected against being run by untrusted users? [00:40:21] awight: yes [00:40:30] ok thx! [00:40:42] description: > [00:40:42] Jobs manually triggered by whitelisted users commenting 'check experimental'. [00:40:42] Useful for experimenting new jobs. [00:42:20] Also: I remember the list of whitelisted users was a serious runon line, but now I don't see it in config at all. Is Gerrit or LDAP responsible for that now? [00:44:29] ah, nvm, found it [00:45:48] awight: it's actually duplicated in two places fyi [00:47:02] argh :) [00:54:43] Project beta-scap-eqiad build #44221: FAILURE in 43 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44221/ [00:58:41] (03PS1) 10Legoktm: Don't ignore l10n-bot [integration/config] - 10https://gerrit.wikimedia.org/r/194990 (https://phabricator.wikimedia.org/T91707) [01:02:38] (03PS1) 10Awight: Expand email regex [integration/config] - 10https://gerrit.wikimedia.org/r/194993 [01:04:04] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1097446 (10Legoktm) [01:04:07] (03PS1) 10Awight: Sort emails [integration/config] - 10https://gerrit.wikimedia.org/r/194994 [01:09:23] 10Continuous-Integration, 10TemplateData: Run JSduck in CI for TemplateData extension - https://phabricator.wikimedia.org/T91851#1097467 (10Jdforrester-WMF) 3NEW [01:15:03] Yippee, build fixed! [01:15:03] Project beta-scap-eqiad build #44223: FIXED in 58 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44223/ [01:19:22] (03PS1) 10Jforrester: Add jsduck and jsduck-publish tasks for TemplateData [integration/config] - 10https://gerrit.wikimedia.org/r/194996 (https://phabricator.wikimedia.org/T91851) [01:34:18] 10Continuous-Integration: Migrate jsduck-publish to use a single job and variable destination path - https://phabricator.wikimedia.org/T91854#1097527 (10Krinkle) p:5Triage>3Low a:3Krinkle [02:46:20] (03CR) 10Krinkle: [C: 032] "Deployed new job: https://integration.wikimedia.org/ci/job/mwext-TemplateData-jsduck-publish/" [integration/config] - 10https://gerrit.wikimedia.org/r/194996 (https://phabricator.wikimedia.org/T91851) (owner: 10Jforrester) [02:46:34] Krinkle: No !log? [02:47:09] James_F: Only for Zuul reload [02:47:31] It's not yet triggered by anything. It's just a lonely xml file on disk now [02:47:32] Well, merging that causes one, doesn't it? [02:47:36] Ah, OK. [02:47:38] Nope [02:47:55] After this is merged, I ssh into gallium, and git pull in /etc/ zuul-config dir , then graceful reload zuul and log that [02:48:05] Sounds fun. [02:49:19] James_F: Unfortunately, Jenkins is now going to compile all 800 jobs from latest master (before merging) and then again (after simulated merge) and show a diff, and then merge. [02:49:25] Which takes 4 minutes each [02:49:29] Why? [02:49:37] Is this just what it does? [02:49:37] because that's the test suite [02:49:40] Why only 800? [02:50:00] There are apparently 3808 output files… [02:50:16] sure, 3800 [02:50:36] <^demon|away> test all the files! [02:50:45] It doesnt' even test it [02:50:50] It just produces a colordiff [02:50:53] of the two directories [02:51:07] That's useful pre-merge, but probably not so much on gate. [02:51:12] Although it could've changed [02:51:21] It makes sense conceptually, it's just very slow. [02:51:24] <^demon|away> Krinkle: Question. Was https://gerrit.wikimedia.org/r/#/c/180118/ due to a problem with having them in the release? [02:51:35] <^demon|away> Or just "tests aren't important for release?" [02:51:38] BUt that'll get faster once we consolidate most extension jobs into one. [02:51:47] 3808*4 minutes = 15232 minutes = 253 hours 52 minutes = 1.51 weeks. [02:51:58] Is it really 4 minutes? [02:51:59] No, 4 minutes for a complete run [02:52:05] Oh, OK. [02:52:06] Youv'e done this locally just now [02:52:18] It does it twice, one for with and wihtout the change [02:52:19] Yeah, 4 minutes sounds about right for how long it took. [02:52:39] (03Merged) 10jenkins-bot: Add jsduck and jsduck-publish tasks for TemplateData [integration/config] - 10https://gerrit.wikimedia.org/r/194996 (https://phabricator.wikimedia.org/T91851) (owner: 10Jforrester) [02:52:55] ^demon|away: Yeah, mostly that. And also to avoid hitting any triggers of it being a Ruby or JS based project. [02:53:07] ^demon|away: I think it was because we'd already made the decision to exclude /test from releases/ [02:53:11] ^demon|away: Which is what I've seen distros like DirectAdmin do [02:53:14] <^demon|away> Those are being added back [02:53:17] Ah. [02:53:26] Why were they removed, anyway? [02:53:46] Why are they being added back, anyway? [02:53:53] That too. :-) [02:54:19] !log Reloading Zuul to deploy Ia82a0d45ac431b5 [02:54:20] <^demon|away> Original patch just says "we don't want them" [02:54:27] Logged the message, Master [02:54:35] <^demon|away> It was brought up on list and in Phab that some people develop against the tarballs for internal usage. [02:54:37] ^demon|away: smaller tarball? [02:54:52] Meh. Whilst I'm feeling CI-y, does someone want to +2 https://gerrit.wikimedia.org/r/195003 so I can deploy it? [02:54:55] <^demon|away> Plus, hooks & docs still make reference to non-existing tests. [02:55:19] <^demon|away> If we want a smaller tarball we should drop i18n. [02:55:31] ^demon|away: Or support for databases. :-) [02:55:46] <^demon|away> Anything other than i18n is nitpicking [02:55:53] Thanks, Krinkle. [02:56:02] ^demon|away: Hm.. Yeah [02:56:05] <^demon|away> It's the only directory in the whole repo that has any size :p [02:56:36] It's quite common for projects to have an 'en' release with 'i18n' drop-in. [02:56:50] E.g. untar on top [02:56:55] Or even pre-made [02:57:02] But gets complicated with upgrades of course. [02:57:04] <^demon|away> Anyway, that's a different discussion :) [02:57:09] Yeah [02:57:54] One could argue for security perhaps. Tests aren't the most tightly reviewed code. WMF doesn't expose many directories for that and other reasons. [02:58:04] Not sure how well we guard that by default [02:58:09] .htaccess isn't everything [02:58:45] Ohwell, we're just making reasons backwards. [02:58:45] <^demon|away> It shouldn't be executable except via the phpunit.php entrypoint really [02:58:54] <^demon|away> (and if it is, people are writing bad tests :p) [02:59:03] I can think of a few more [02:59:10] * Krinkle wrote them [02:59:14] <^demon|away> hehe [02:59:39] https://github.com/wikimedia/mediawiki/blob/master/tests/qunit/data/load.mock.php [02:59:48] https://github.com/wikimedia/mediawiki/blob/master/tests/qunit/data/styleTest.css.php [02:59:52] <^demon|away> Use Phab links! [02:59:52] !log Pushed Ib4f6e9 and Ie26bb17 to grrrit-wm and restarted [02:59:56] Nice echo servers for arbitrary content [02:59:56] Logged the message, Master [03:00:27] ^demon|away: It doesn't sort directories above files [03:00:34] 2 seconds lost of my life [03:00:52] <^demon|away> lolol [03:00:53] https://phabricator.wikimedia.org/diffusion/MW/browse/master/tests/qunit/data/styleTest.css.php [03:01:01] No thanks [03:01:22] Krinkle: {{sofixit}}. ;-) [03:01:26] I can't read that without glasses. [03:01:50] s/can't/don't want to/ [03:02:04] <^demon|away> I need to finish the last repos [03:02:09] ^demon|away: Yes! [03:02:17] So we can close that Phabricator ticket and rejoice. [03:02:20] <^demon|away> Other than the Phab repos cuz that makes shit happen [03:02:23] And then burn gitblit to the ground. [03:02:34] Redirects! [03:02:53] Sure. [03:02:56] That's a domain issue. [03:03:02] But burn *gitblit* to the ground. [03:03:17] And presumably a lot more because we don't use it as viewer, but also as mirror. Although maybe the mirror will stay? [03:03:22] Lots of stuff references git.wikimedia.org [03:03:25] It's a nice domain name [03:03:37] I think ^demon|away got the GitHub replication etc. working from it. [03:03:47] <^demon|away> Not yet [03:04:31] Ah. Shame. [03:04:46] James_F: btw, I see Jenkins is still updating mediawiki/extensions.git every commit [03:04:47] for VE [03:04:47] Krinkle: Yup. Isn't gerrit great? [03:04:53] What about all other repos [03:04:59] Why doesn't auto work for VE? [03:05:31] Because we have two repos called VisualEditor inside each other. [03:05:37] And gerrit can't cope with that. [03:05:43] Or something. ^demon|away knows the details. [03:05:53] That would explain VisualEditor/VisualEditor [03:05:58] This is mwext though [03:06:03] Oh. [03:06:12] <^demon|away> It's confusing and weird [03:06:24] No, the VisualEditor/VisualEditor repo poisons gerrit so /VisualEditor anywhere else doesn't work, or something. [03:06:33] <^demon|away> It doesn't like multiple things named */VisualEditor yeah [03:06:44] ^demon|away: And */core ? [03:06:59] <^demon|away> If you had auto-updating repo thingies yeah [03:07:02] <^demon|away> It's just that [03:07:05] Krinkle: We don't have any auto– yeah. [03:07:11] oojs-core / mediawiki-core [03:07:12] Yeah [03:07:33] If only we'd ignored ^demon|away and called it VisualEditor/core it'd have been fine! ;-P [03:07:47] <^demon|away> Or just called it VisualEditor? [03:07:52] Indeed [03:07:57] We're planning to have further VE repos. [03:07:58] <^demon|away> /core has been redundant and stupid [03:07:58] Like unicodejs [03:08:06] VisualEditor-WordPress [03:08:09] VisualEditor-Drupal [03:08:12] James_F: They can use dashes [03:08:16] <^demon|away> Anyway, the bug with repo names and auto-update is batshit. James_F and I spent like 3-4 days trying to figure it out and we finally gave up. [03:08:16] VisualEditor-Etc. [03:08:17] Who needs namespacing [03:08:20] Can. Whatever. [03:08:28] All moot now. [03:08:33] All hail Phabricator [03:08:34] It's what we'll have in Phabricator I guess [03:08:44] I wonder how its git server is exposed. [03:08:50] Yup, snappily-named QQZD repos. :-) [03:08:52] Right now the Phab repos point to gerrit in the clone instructions [03:08:54] <^demon|away> You can expose it over https or ssh [03:08:58] <^demon|away> Right now we have neither. [03:09:04] <^demon|away> chase and I haven't poked it yet [03:09:10] Soon, soon! [03:09:14] <^demon|away> We'll probably just do https. ssh is a pain in the ass. [03:09:16] But after we're done. [03:09:24] For read-only mirror, HTTPS is enough. [03:09:40] https://phabricator.wikimedia.org/diffusion/MW/history/master/ [03:09:53] Not bad [03:10:04] Except the most important column is the second-tinyiest [03:10:15] Which one? [03:10:19] Details [03:10:22] Narrow window [03:10:34] It's the largest column for me. [03:10:40] Its what happens when you try to squish it into a table [03:10:45] Over 1800px wide. [03:10:51] Make it 900px [03:11:01] it'll be narrower than the date column [03:11:04] Still the widest column. [03:11:05] Nope. [03:11:25] Shortening "Gerrit Code Review" would help a lot, though. [03:11:28] http://i.imgur.com/IhHvbcW.png [03:11:39] Krinkle: That is poor. :-( [03:11:49] Solution: Get a better monitor. [03:11:57] No, I made it a narrow window [03:12:04] My resolution is equal to yours [03:12:12] higher actually, but I won't say the word Retina. [03:12:21] Sure. [03:12:38] https://github.com/wikimedia/mediawiki/commits/master [03:12:47] Yuck. [03:12:49] Two-row view. [03:12:53] These two urls are classic developer vs. designer [03:12:56] Hate hate hate. [03:13:09] Yeah, GitHub is all pretty and no value. [03:13:25] well designed, not just any designer [03:14:02] https://phabricator.wikimedia.org/rMW3702223dda048bcf17e5b651be3d81145d5315c2 is rather swanky. [03:14:03] I think it's the kind of case where literally every person on earth will come out being able to use other interface faster and more intuitive. Even if their gut doesn't like it. [03:14:30] The git hash references are seemlessly turned into links across repos. [03:14:53] That's nice [03:15:11] And the single-page all-files diff is particularly nice. [03:15:19] So you can compare changes between files at a glance. [03:15:24] Yeah [03:15:29] Though GitHub's had that since day 1 [03:15:36] Gerrit hasn't. :-) [03:15:41] Who uses that [03:15:42] :P [03:15:43] Which is what we're comparing to, remember. [03:15:55] https://phabricator.wikimedia.org/rCICF487e918593902dcbc065bc7ba490323f15bffb9c similarly. [03:15:55] Yeah [03:16:19] James_F: Hm.. seems nowhere on the page is the raw git hash [03:16:28] In a selectable fashion [03:16:31] Krinkle: It's in the title. [03:16:35] Only with the rCASFSDF prefix [03:16:50] rCICF [03:16:51] But yes. [03:18:02] I don't get why all of diffusion has awesome urls and then a squishes single string for the diff [03:18:04] https://phabricator.wikimedia.org/diffusion/CICF/repository/master/ [03:18:13] https://phabricator.wikimedia.org/diffusion/CICF/repository/commit/487e918593902dcbc065bc7ba490323f15bffb9c [03:18:15] or something like that [03:19:13] URL hacking is variably loved and loathed. [03:19:15] * James_F shrugs. [03:19:16] Anyway. [03:19:21] or at least a dash or slash so it can be selected [03:19:23] I should consider getting food and going home. [03:19:31] And I now what I should [03:19:35] And Krinkle: write me code or go away. :-) [03:19:39] know* [03:19:50] <^demon|away> James_F: I gave you the list of gerrit repos not in Phab. Could you massage that list into the on-wiki list? [03:19:50] * James_F grins. [03:20:09] <^demon|away> (and just drop any rows that are {{done}}) [03:20:17] ^demon|away: Oh, sure. Did you give it to me in IRC? My client crashed this morning with the DHCP oddities. :-( [03:20:22] <^demon|away> On Phab [03:20:34] I'm meant to read the e-mails from Phab?! [03:20:35] ;-) [03:20:43] <^demon|away> https://phabricator.wikimedia.org/P349 [03:21:01] * James_F updates. [03:21:41] Project browsertests-CirrusSearch-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce build #492: FAILURE in 1 min 40 sec: https://integration.wikimedia.org/ci/job/browsertests-CirrusSearch-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce/492/ [03:21:54] <^demon|away> Ok, I'm gonna get food and go to big kid toys-r-us [03:22:10] <^demon|away> Have a good weekend gentlemen [03:22:23] And you. [03:25:43] (03PS1) 10Krinkle: TemplateData: Remove redundant 'jslint' and 'lint' jobs from test pipeline [integration/config] - 10https://gerrit.wikimedia.org/r/195006 [03:27:44] (03PS2) 10Krinkle: TemplateData: Remove redundant 'jslint' and 'lint' jobs from test pipeline [integration/config] - 10https://gerrit.wikimedia.org/r/195006 [03:28:17] (03PS3) 10Krinkle: TemplateData: Remove redundant 'jslint' and 'lint' jobs from test pipeline [integration/config] - 10https://gerrit.wikimedia.org/r/195006 [03:30:10] 10Continuous-Integration, 10TemplateData: Run JSduck in CI for TemplateData extension - https://phabricator.wikimedia.org/T91851#1097612 (10Krinkle) 5Open>3Resolved a:3Krinkle [03:30:20] 10Continuous-Integration, 10TemplateData: Run JSDuck in CI for TemplateData extension - https://phabricator.wikimedia.org/T91851#1097617 (10Krinkle) [03:31:10] (03CR) 10Krinkle: [C: 032] TemplateData: Remove redundant 'jslint' and 'lint' jobs from test pipeline [integration/config] - 10https://gerrit.wikimedia.org/r/195006 (owner: 10Krinkle) [03:31:22] !log Reloading Zuul to deploy I30131a32c7f1 [03:31:26] Logged the message, Master [03:32:14] (03Merged) 10jenkins-bot: TemplateData: Remove redundant 'jslint' and 'lint' jobs from test pipeline [integration/config] - 10https://gerrit.wikimedia.org/r/195006 (owner: 10Krinkle) [03:34:36] Krinkle: Queue-jumper. :-) [03:35:17] James_F: Nah, my IRC process is asynchronous I logged that before I started ssh-ing into gallium [03:35:28] Oh, right. [03:35:46] I just figured I'd get that over with while it's busy merging [03:36:01] * James_F nods. [03:44:21] (03PS1) 10Krinkle: Replace 'mwext-VisualEditor-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195007 [03:45:19] James_F: There's probably about 400 replacements like these ^ to be done for other stuff. Step by step. It's a nice thing to do in the background while relaxing. It literally is most just removing code, and repeatedly waiting 2x (test/merge) 2x (before/after) 4 minutes [03:46:32] Which in total will drastically cut down our disk usage and prepare us for isolation. [03:46:57] And also make those 4 minutes much shorter a there shouldn't be 500 -lint jobs. [03:47:24] All those macros, finally paying off [03:47:29] Its just slashing away now [03:50:07] Krinkle: :-) [03:50:17] Krinkle: Not tempted to kill 50 of them in one go? [03:50:33] James_F: I will, I'm doing smaller ones first. [03:50:43] For phplint, we'll delete 900 in one go [03:50:52] but there is only 12 ducks [03:51:12] Fun. [03:51:13] Wanna make sure the new infra is tested gradually. [03:51:26] already discovered a few minor bugs, but nothing serious. [03:52:36] (03CR) 10Krinkle: [C: 032] Replace 'mwext-VisualEditor-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195007 (owner: 10Krinkle) [03:53:06] (03PS3) 10Krinkle: oojs-ui-jsduck-publish: Run build to support @example demos [integration/config] - 10https://gerrit.wikimedia.org/r/194957 [03:53:14] (03PS4) 10Krinkle: CentralNotice: enable qunit-karma job [integration/config] - 10https://gerrit.wikimedia.org/r/194773 (https://phabricator.wikimedia.org/T86092) (owner: 10AndyRussG) [03:55:45] !log Pool integration-slave1402 [03:55:50] Logged the message, Master [03:57:16] awight: Hi. How'd go? [03:58:58] (03Merged) 10jenkins-bot: Replace 'mwext-VisualEditor-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195007 (owner: 10Krinkle) [04:00:04] (03PS4) 10Krinkle: oojs-ui-jsduck-publish: Run build to support @example demos [integration/config] - 10https://gerrit.wikimedia.org/r/194957 [04:02:37] Krinkle: very well, but I got cocky and ignored your advice to create a m1.medium... I think I hit an OOM on the m1.small, and the mysql-server was corrupted. So I'm reprovisioning now, as a medium... [04:02:53] OK [04:03:10] I noticed the jenkins node entry showing an error when I added slave1402 [04:03:19] for civicrm not existing [04:03:20] :) [04:03:22] That explains [04:03:34] No worries, I was just nosing around [04:03:59] :D thanks for all the help! [04:04:35] (03PS1) 10Krinkle: Replace 'mwext-MultimediaViewer-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195008 [04:08:03] (03PS1) 10Krinkle: macro: Remove unused 'jsduck-conf' [integration/config] - 10https://gerrit.wikimedia.org/r/195009 [04:09:34] ^demon|away: When you get back, https://www.mediawiki.org/wiki/Phabricator/Diffusion/Callsign_naming_conventions/Existing_repositories now has all {{done}} repos removed (and a bunch I noticed were done but weren't marked as such – there are likely some I've skipped), and all the remaining ones that aren't obvious non-transfers added (mostly new repos). [04:10:05] (03CR) 10Jforrester: macro: Remove unused 'jsduck-conf' (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/195009 (owner: 10Krinkle) [04:10:17] And with that, I actually will go. [04:25:03] Project browsertests-UploadWizard-commons.wikimedia.beta.wmflabs.org-linux-firefox-sauce build #525: FAILURE in 28 min: https://integration.wikimedia.org/ci/job/browsertests-UploadWizard-commons.wikimedia.beta.wmflabs.org-linux-firefox-sauce/525/ [04:28:47] (03CR) 10Krinkle: macro: Remove unused 'jsduck-conf' (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/195009 (owner: 10Krinkle) [04:33:45] (03CR) 10Krinkle: [C: 032] "Deployed oojs-ui-jsduck-publish." [integration/config] - 10https://gerrit.wikimedia.org/r/194957 (owner: 10Krinkle) [04:40:11] (03Merged) 10jenkins-bot: oojs-ui-jsduck-publish: Run build to support @example demos [integration/config] - 10https://gerrit.wikimedia.org/r/194957 (owner: 10Krinkle) [04:40:37] Project browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-chrome-sauce build #558: STILL FAILING in 19 min: https://integration.wikimedia.org/ci/job/browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-chrome-sauce/558/ [04:45:19] (03PS2) 10Krinkle: Replace 'mwext-MultimediaViewer-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195008 [04:45:23] (03CR) 10Krinkle: [C: 032] Replace 'mwext-MultimediaViewer-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195008 (owner: 10Krinkle) [04:48:46] 10Continuous-Integration: Jenkins: Assert no PHP errors (notices, warnings) were raised or exceptions were thrown - https://phabricator.wikimedia.org/T50002#1097658 (10Krinkle) [04:50:49] (03CR) 10Krinkle: "I can't see anything out of the ordinary compared to mediawiki-core and mwext-VisualEditor's qunit-karma runs. I re-ran it twice and the s" [integration/config] - 10https://gerrit.wikimedia.org/r/194773 (https://phabricator.wikimedia.org/T86092) (owner: 10AndyRussG) [04:51:17] !log Pool integration-slave1403 [04:51:21] Logged the message, Master [04:51:34] (03Merged) 10jenkins-bot: Replace 'mwext-MultimediaViewer-jsduck with 'jsduck' [integration/config] - 10https://gerrit.wikimedia.org/r/195008 (owner: 10Krinkle) [04:52:39] (03PS2) 10Krinkle: macro: Remove unused 'jsduck-conf' [integration/config] - 10https://gerrit.wikimedia.org/r/195009 [04:52:44] (03CR) 10Krinkle: [C: 032] macro: Remove unused 'jsduck-conf' [integration/config] - 10https://gerrit.wikimedia.org/r/195009 (owner: 10Krinkle) [04:53:11] !log Reloading Zuul to deploy I585b7f026 [04:53:15] Logged the message, Master [04:58:22] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1097662 (10Krinkle) I fully agree we should make it run the tests. I'm concerned about the non-trivial stress this might put on the c... [04:58:51] (03Merged) 10jenkins-bot: macro: Remove unused 'jsduck-conf' [integration/config] - 10https://gerrit.wikimedia.org/r/195009 (owner: 10Krinkle) [04:58:53] (03CR) 10Krinkle: [C: 04-1] "Per T91707, needs a quick research with regards to CI impact." [integration/config] - 10https://gerrit.wikimedia.org/r/194990 (https://phabricator.wikimedia.org/T91707) (owner: 10Legoktm) [05:00:41] 10Continuous-Integration, 5Patch-For-Review: Jenkins: point TMPDIR to dir in workspace or tmpfs and delete after build - https://phabricator.wikimedia.org/T70563#1097664 (10Krinkle) [05:00:42] 10Continuous-Integration, 6MediaWiki-Core-Team: /tmp/mw-Parser-*/thumb/*.jpg files left behind on Jenkins slaves in production - https://phabricator.wikimedia.org/T91209#1097663 (10Krinkle) [05:01:04] 10Continuous-Integration, 6MediaWiki-Core-Team: /tmp/mw-Parser-*/thumb/*.jpg files left behind on Jenkins slaves in production - https://phabricator.wikimedia.org/T91209#1097665 (10Krinkle) 5Open>3Resolved a:3Krinkle > mw-install-sqlite: Expose tmpfs dir as TMPDIR > https://gerrit.wikimedia.org/r/194038 [05:01:05] 10Continuous-Integration: Jenkins: Figure out long term solution for /tmp management - https://phabricator.wikimedia.org/T74011#1097668 (10Krinkle) [05:01:19] 10Continuous-Integration: Jenkins: Figure out long term solution for /tmp management - https://phabricator.wikimedia.org/T74011#759885 (10Krinkle) [05:01:20] 10Continuous-Integration, 5Patch-For-Review: Jenkins: point TMPDIR to dir in workspace or tmpfs and delete after build - https://phabricator.wikimedia.org/T70563#727272 (10Krinkle) [05:49:18] (03PS1) 10Krinkle: Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 [05:55:18] (03CR) 10jenkins-bot: [V: 04-1] Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 (owner: 10Krinkle) [05:56:24] (03PS2) 10Krinkle: Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 [06:02:30] (03CR) 10jenkins-bot: [V: 04-1] Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 (owner: 10Krinkle) [06:08:10] (03PS3) 10Krinkle: Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 [06:15:01] (03CR) 10Krinkle: [C: 032] "Deployed 'phplint' job." [integration/config] - 10https://gerrit.wikimedia.org/r/195018 (owner: 10Krinkle) [06:21:07] (03CR) 10jenkins-bot: [V: 04-1] Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 (owner: 10Krinkle) [06:21:15] (03PS4) 10Krinkle: Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 [06:21:22] (03CR) 10Krinkle: [C: 032] Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 (owner: 10Krinkle) [06:27:35] (03Merged) 10jenkins-bot: Consolidate all *-lint and *-phplint jobs into a generic 'phplint job [integration/config] - 10https://gerrit.wikimedia.org/r/195018 (owner: 10Krinkle) [06:28:03] !log Reloading Zuul to deploy I8d7e0bd315c4fc2 [06:28:09] Logged the message, Master [06:32:52] (03PS1) 10Krinkle: Remove unused csslint templates [integration/config] - 10https://gerrit.wikimedia.org/r/195019 [06:34:04] (03PS1) 10Krinkle: macro: Remove unused 'jshint-in' [integration/config] - 10https://gerrit.wikimedia.org/r/195020 [06:35:12] RECOVERY - Free space - all mounts on deployment-bastion is OK: OK: All targets OK [06:35:55] PROBLEM - Puppet failure on deployment-salt is CRITICAL: CRITICAL: 33.33% of data above the critical threshold [0.0] [06:41:05] (03PS1) 10Krinkle: scm: Enable shallow-clone for git-remoteonly-zuul [integration/config] - 10https://gerrit.wikimedia.org/r/195021 [06:41:27] (03CR) 10Krinkle: [C: 032] Remove unused csslint templates [integration/config] - 10https://gerrit.wikimedia.org/r/195019 (owner: 10Krinkle) [06:41:50] (03CR) 10Krinkle: [C: 032] macro: Remove unused 'jshint-in' [integration/config] - 10https://gerrit.wikimedia.org/r/195020 (owner: 10Krinkle) [06:47:01] (03Merged) 10jenkins-bot: Remove unused csslint templates [integration/config] - 10https://gerrit.wikimedia.org/r/195019 (owner: 10Krinkle) [06:47:42] (03Merged) 10jenkins-bot: macro: Remove unused 'jshint-in' [integration/config] - 10https://gerrit.wikimedia.org/r/195020 (owner: 10Krinkle) [06:54:20] (03PS1) 10Krinkle: Consolidate pplint-HEAD and erblint-HEAD jobs into generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/195022 [06:55:37] (03PS2) 10Krinkle: Consolidate pplint-HEAD and erblint-HEAD jobs into generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/195022 [07:00:55] RECOVERY - Puppet failure on deployment-salt is OK: OK: Less than 1.00% above the threshold [0.0] [07:01:06] (03CR) 10jenkins-bot: [V: 04-1] Consolidate pplint-HEAD and erblint-HEAD jobs into generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/195022 (owner: 10Krinkle) [07:05:20] (03PS3) 10Krinkle: Consolidate pplint-HEAD and erblint-HEAD jobs into generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/195022 [07:10:29] (03PS4) 10Krinkle: Consolidate pplint-HEAD and erblint-HEAD jobs into generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/195022 [07:14:37] Project beta-scap-eqiad build #44259: FAILURE in 40 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44259/ [07:34:58] Yippee, build fixed! [07:34:58] Project beta-scap-eqiad build #44261: FIXED in 57 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44261/ [07:56:33] PROBLEM - Puppet failure on deployment-pdf01 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [0.0] [08:24:16] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1097793 (10Legoktm) For reference, in https://gerrit.wikimedia.org/r/#/c/102636/ @hashar blacklisted l10n-bot. As for how many repos... [08:41:36] https://github.com/JakubOnderka/PHP-Parallel-Lint/issues/47 [09:00:47] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1097808 (10Nemo_bis) It might be easier to just revert 02364f56df90ca9e64a3d63fd81c52643e4f9e97 : Raimond used to act on test failures... [09:02:07] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1097810 (10Nemo_bis) p:5Triage>3Normal [10:57:49] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1097872 (10Raymond) Would it help to change the time when I do the export from translatewiki.net? Currently I export and submit nearly... [12:37:35] PROBLEM - SSH on deployment-lucid-salt is CRITICAL: Connection refused [13:34:46] Project beta-scap-eqiad build #44297: FAILURE in 43 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44297/ [13:54:35] 6Release-Engineering, 6MediaWiki-Core-Team, 6Multimedia, 6Parsoid-Team, and 3 others: Prepare Platform April 2015 quarterly review presentation - https://phabricator.wikimedia.org/T91803#1098059 (10Qgil) [13:54:54] Yippee, build fixed! [13:54:54] Project beta-scap-eqiad build #44299: FIXED in 51 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44299/ [14:14:15] (03CR) 10Krinkle: [C: 032] "Deployed new pplint-HEAD and erblint-HEAD jobs." [integration/config] - 10https://gerrit.wikimedia.org/r/195022 (owner: 10Krinkle) [14:28:47] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1098074 (10Krinkle) Hm.. If it's automated and could be any time of day, then perhaps between 03:00 and 06:00 UTC. That would match 04... [14:32:33] (03Merged) 10jenkins-bot: Consolidate pplint-HEAD and erblint-HEAD jobs into generic jobs [integration/config] - 10https://gerrit.wikimedia.org/r/195022 (owner: 10Krinkle) [14:35:17] !log Reloading Zuul to deploy I864875aa4acc [14:35:23] Logged the message, Master [14:36:54] Project browsertests-MobileFrontend-SmokeTests-linux-chrome-sauce build #43: STILL FAILING in 8 min 53 sec: https://integration.wikimedia.org/ci/job/browsertests-MobileFrontend-SmokeTests-linux-chrome-sauce/43/ [14:41:47] !log Pool integration-slave1404 [14:41:51] Logged the message, Master [14:43:49] !log Depool integration-slave1006 and integration-slave1007 (not deleting yet, just in case) [14:43:52] Logged the message, Master [14:44:23] !log Depool integration-slave1008 and integration-slave1010 (not deleting yet, just in case) [14:44:26] Logged the message, Master [14:55:47] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1098102 (10hashar) I haven't looked back at the l10n-bot configuration in Zuul, but what could be done is: * blacklist bot from test... [15:00:20] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1098120 (10Krinkle) Note that the extra time spent in CI (~ 5 minutes) is enough to occasionally cause race conditions where, after th... [15:33:28] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1098149 (10Nemo_bis) The export is not fully automated, no. [15:47:58] 10Continuous-Integration, 6translatewiki.net, 5Patch-For-Review: l10n-bot self-force-merging sometimes breaks mediawiki/core master - https://phabricator.wikimedia.org/T91707#1098158 (10Raymond) >>! In T91707#1098074, @Krinkle wrote: > Or between [16:00 and 19:00 UTC](http://everytimezone.com/#2015-3-7,240,c... [16:31:32] (03PS1) 10Awight: Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 [16:32:39] (03CR) 10jenkins-bot: [V: 04-1] Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (owner: 10Awight) [16:33:04] (03PS2) 10Awight: Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) [16:34:08] (03CR) 10jenkins-bot: [V: 04-1] Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [16:35:09] There's some kind of blessing I have to give my jobs, I guess? Maybe they have to come from JJB rather than being hand-coded directly in the Jenkins console? FAIL: test_valid_jobs_in_check_pipelines (test_zuul_layout.TestZuulLayout) [16:36:12] (03CR) 10Awight: "Looks like I need to make a JJB definition for this job?" [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [16:47:03] (03CR) 10Jforrester: macro: Remove unused 'jsduck-conf' (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/195009 (owner: 10Krinkle) [16:55:07] 10Continuous-Integration: mediawiki-extensions-hhvm failed on a patchset with "RuntimeException: Cannot override frozen service "storage"." - https://phabricator.wikimedia.org/T91888#1098226 (10Jdforrester-WMF) 3NEW [17:30:01] PROBLEM - Long lived cherry-picks on puppetmaster on deployment-salt is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [0.0] [18:35:06] Yippee, build fixed! [18:35:06] Project browsertests-CirrusSearch-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce build #493: FIXED in 2 min 6 sec: https://integration.wikimedia.org/ci/job/browsertests-CirrusSearch-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce/493/ [19:11:10] 10Continuous-Integration, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM: Write Jenkins job builder definition for CiviCRM CI job - https://phabricator.wikimedia.org/T91895#1098353 (10awight) 3NEW a:3awight [19:36:32] (03CR) 10Hashar: "I am happy to see this change coming back. Thank you!" (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/194993 (owner: 10Awight) [19:36:35] (03PS2) 10Hashar: Expand email regex [integration/config] - 10https://gerrit.wikimedia.org/r/194993 (owner: 10Awight) [19:36:40] (03CR) 10Hashar: [C: 032] Expand email regex [integration/config] - 10https://gerrit.wikimedia.org/r/194993 (owner: 10Awight) [19:36:58] (03PS2) 10Hashar: Sort emails [integration/config] - 10https://gerrit.wikimedia.org/r/194994 (owner: 10Awight) [19:37:31] (03CR) 10Hashar: [C: 032] Sort emails [integration/config] - 10https://gerrit.wikimedia.org/r/194994 (owner: 10Awight) [19:37:45] Yippee, build fixed! [19:37:45] (03Merged) 10jenkins-bot: Expand email regex [integration/config] - 10https://gerrit.wikimedia.org/r/194993 (owner: 10Awight) [19:37:45] Project browsertests-UploadWizard-commons.wikimedia.beta.wmflabs.org-linux-firefox-sauce build #526: FIXED in 34 min: https://integration.wikimedia.org/ci/job/browsertests-UploadWizard-commons.wikimedia.beta.wmflabs.org-linux-firefox-sauce/526/ [19:38:48] (03Merged) 10jenkins-bot: Sort emails [integration/config] - 10https://gerrit.wikimedia.org/r/194994 (owner: 10Awight) [19:44:00] PROBLEM - Puppet failure on deployment-cache-mobile03 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [0.0] [19:44:26] (03CR) 10Hashar: [C: 04-1] "The failing test makes sure that jobs in 'check-voter' and 'check' pipelines are only whitelisted jobs. We do not want to run jobs execut" (032 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [19:52:34] (03PS5) 10Hashar: Package python deps with dh-virtualenv [integration/zuul] (debian-precise-venv) - 10https://gerrit.wikimedia.org/r/194520 [19:53:01] (03CR) 10Hashar: "check experimental" [integration/zuul] (debian-precise-venv) - 10https://gerrit.wikimedia.org/r/194520 (owner: 10Hashar) [19:53:28] (03PS3) 10Awight: Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) [19:55:08] (03CR) 10Awight: "Thank you for the explanation!" (032 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [19:55:31] Project beta-scap-eqiad build #44334: FAILURE in 46 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44334/ [20:14:28] (03PS1) 10Awight: whitespace and quotes fixup [integration/config] - 10https://gerrit.wikimedia.org/r/195061 [20:15:00] Yippee, build fixed! [20:15:01] Project beta-scap-eqiad build #44336: FIXED in 1 min 2 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/44336/ [20:16:54] (03PS1) 10Awight: Run a new experimental job for CRM patches [integration/config] - 10https://gerrit.wikimedia.org/r/195062 (https://phabricator.wikimedia.org/T91895) [20:17:56] (03PS1) 10Awight: WIP: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [20:23:14] (03Abandoned) 10Awight: Run a new experimental job for CRM patches [integration/config] - 10https://gerrit.wikimedia.org/r/195062 (https://phabricator.wikimedia.org/T91895) (owner: 10Awight) [20:32:49] hashar: Oops, I just ran an unsafe job on the wrong slave, https://integration.wikimedia.org/ci/job/wikimedia-fundraising-civicrm/1/consoleFull [20:33:06] awight: hey :) [20:33:09] good to see progress! [20:34:15] awight: the jobs can run anywhere [20:34:27] awight: it is just that we have a whitelist of people in zuul/layout.yamll [20:34:41] so we trigger the jobs only for whitelisted people [20:34:45] Hehe, it turns out the job will run well on a standard ci slave. But it's probably wiser to keep it isolated, because it downloads packages via composer... [20:34:58] It also makes a mess in MySQL, but cleans up afterwards. [20:35:25] awight: just tie it to labs slave [20:36:10] hashar: That's okay, that it downloads the whole world? [20:36:10] awight: in JJB job definition node: contintLabsSlave && UbuntuTrusty [20:36:21] awight: not really :) [20:36:31] awight: but we have ton of jobs already downloading the whole world on labs instance [20:36:34] such as composer [20:36:37] or npm [20:36:50] Ah, okay I'll join the party then :) [20:38:39] Question: How does the phpunit-junit builder determine which directory to test? Is there a phpunit.xml in $WORKSPACE? [20:38:54] hold on [20:38:58] online shopping going on :) [20:39:07] lol, bidsniper [20:40:04] done [20:40:10] kid need a new coat [20:40:19] we settled with my wife that she will be offered to choose the color [20:40:29] Constrained choice :p [20:40:44] yeah gray, red or yellow [20:40:47] my bet is on yellow [20:41:03] so [20:41:04] +1 [20:41:08] the job is using zuul-cloner [20:41:14] at the top of https://integration.wikimedia.org/ci/job/wikimedia-fundraising-civicrm/1/consoleFull [20:41:21] yes [20:41:24] there is some colored info/debug messages emitted by zuul cloner [20:41:49] it should show the destination path for each repo [20:42:07] Mapping projects to workspace... [20:42:07] wikimedia/fundraising/crm -> src/wikimedia/fundraising/crm [20:42:07] wikimedia/fundraising/civicrm-buildkit -> src/wikimedia/fundraising/civicrm-buildkit [20:42:33] afaik, everything was healthy until 20:31:58 + /srv/deployment/integration/phpunit/vendor/bin/phpunit --log-junit log/junit-phpunit.xml [20:42:47] ahh [20:42:58] I tried to borrow that builder, and I don't understand how it finds phpunit.xml without a cwd. [20:43:06] yeah so your job populates the db then run composer which download phpunit [20:43:18] but in your JJB job definition you seem to be using a macro we are using for mediawiki/core [20:43:21] yes [20:43:24] which uses a git-deploy version of phpunit [20:43:26] Yippee, build fixed! [20:43:27] Project browsertests-UniversalLanguageSelector-commons.wikimedia.beta.wmflabs.org-linux-firefox-sauce build #503: FIXED in 14 min: https://integration.wikimedia.org/ci/job/browsertests-UniversalLanguageSelector-commons.wikimedia.beta.wmflabs.org-linux-firefox-sauce/503/ [20:43:30] so you will want to come up with your own command [20:43:39] ok, makes sense! [20:44:03] or maybe the repo wikimedia-fundraising-crm can provide a test entry point ? [20:44:06] ie: composer test [20:44:13] that will take care of running phpunit [20:44:16] Still, what is controlling the cwd? I'm interested in getting that junit report, if possible [20:44:26] let me find out [20:44:38] we have more or less a convention [20:45:22] awight: http://www.mediawiki.org/wiki/Continuous_integration/Test_entry_points#PHP [20:45:50] the trick is to define the composer 'test' command with env variables [20:45:54] such as: "phpunit $PHPUNIT_ARGS" [20:46:16] great, that looks nice [20:46:25] this way that would one day let us inject additional args [20:46:47] so Jenkins would ideally run: PHPUNIT_ARGS="--log-junit $WORKSPACE/log/junit-phpunit.xml" composer test [20:47:01] thus the command you defined: phpunit $PHPUNIT_ARGS [20:47:13] will turn out in Jenkins to: phpunit --log-junit $WORKSPACE/log/junit-phpunit.xml [20:47:35] and thus we can add in JJB the publisher macro that handles the phpunit xml file and craft some kind of report in the Jenkins web GUI [20:47:40] I hope it makes sense [20:47:51] It does, thanks [20:49:11] we need to update the JJB templates though :( [20:49:29] but you can already adopt that convention for your job [20:49:51] (03PS2) 10Awight: WIP: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [20:50:49] I'll create a TODO to clean up the test entry point. Currently, it's simply 'phpunit' in the top-level directory, which my team is trained to use :) [20:57:22] 10Continuous-Integration, 6Release-Engineering, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, and 2 others: Create and provision CI slave instance for CiviCRM testing - https://phabricator.wikimedia.org/T89894#1098481 (10awight) It turns out, we don't need our own slave for this job. We can... [20:59:48] (03PS3) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [21:09:14] (03PS4) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [21:11:32] awight: you can refresh the job from your git working copy [21:11:40] and send the yaml change in Gerrit once you are happy with it :) [21:12:03] (though sending it to Gerrit triggers the jobs that might be helpful) [21:13:31] 10Continuous-Integration, 10Wikimedia-Fundraising-CiviCRM: Make Civi CI job run on civicrm, drupal, and vendor repos - https://phabricator.wikimedia.org/T91905#1098495 (10awight) 3NEW a:3awight [21:14:08] 10Continuous-Integration, 10Wikimedia-Fundraising-CiviCRM: Make Civi CI job run on civicrm, drupal, and vendor (DonationInterface and SmashPig) repos - https://phabricator.wikimedia.org/T91905#1098503 (10awight) [21:14:40] (03PS4) 10Awight: Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) [21:15:18] hashar: I've used jjb to push the job and I'm happy with the results, but now I need to add the zuul triggers... [21:16:01] awight: I think I reviewed the change a couple hours ago [21:16:20] This is a new one, then... https://gerrit.wikimedia.org/r/195063 [21:17:28] (03PS5) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [21:18:38] I'm afraid I don't understand these zuul lines, so I've just copy-pastad, https://gerrit.wikimedia.org/r/#/c/195063/4..5/jjb/wm-fundraising.yaml,unified [21:18:52] yeah it is messy :D [21:19:49] It's pretty nice compared to the alternative! I can imagine some places just dump the actual job XML to their VCS repo... [21:20:47] awight: I am reviewing your jjb change [21:21:04] thx [21:23:04] I'm not actually sure I want to use defaults: use-remote-zuul [21:23:20] na you can probably drop it [21:23:31] it is used to set up the Jenkins Git Plugin [21:23:37] k [21:23:40] but we are slowly migrating to zuul-cloner [21:23:52] (which is really a python wrapper invoking the CLI git clone) [21:23:57] (03PS6) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [21:24:05] Yeah, cos I'm already calling zuul-cloner directly. [21:24:51] I have an interesting TODO after this is working... I'd like to run the same job for changes in the submodules as well, but I'll have to fix how the cloning happens in order to do that. [21:29:06] (03CR) 10Hashar: Jenkins job builder definition for CRM job (035 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) (owner: 10Awight) [21:29:34] I suppose a decent workaround would be to run the job on the submodule bump in the parent repo. [21:31:04] yeah that is why zuul-cloner does not process submodules [21:31:16] you are supposed to clone each of them with zuul-cloner [21:31:24] but for a first pass that is already a good thing [21:31:44] if you get all your repositories cloned by zuul-cloner, you can then add Gerrit trigger the same job for all those repositories [21:32:31] cool, cos it will default to master if the $ZUUL_CHANGE doesn't exist in the repo [21:32:38] there is some rather lenghty / nice explanation at http://ci.openstack.org/zuul/gating.html [21:32:45] in short: [21:32:51] if you have repos A and B [21:32:54] That's a little bit of a pain though, it means we need to edit the CI job whenever we add a submodule. [21:33:00] you send Change 1 on repo A and Change 2 on repo B [21:33:04] vote CR+2 on each change [21:33:17] the job can zuul-cloner repo A and repo B [21:33:34] the first build will be using repo A  @change 1 and repo B @ tip of branch [21:33:53] the second build will be using repo A @change 1 and repo B @change2 [21:34:01] ie change2 is tested as if change1 got merged [21:34:29] that's if B is a submodule of A? [21:34:47] * awight reads the link [21:35:16] awight: there is no submodule involved [21:35:42] I wrote the second part of the doc (Cross projects dependencies) on the page [21:35:49] but it is not straightforward [21:38:10] awight: ah yeah https://gerrit.wikimedia.org/r/#/c/195042/4/zuul/layout.yaml [21:38:19] notice how the same job was on all repositories [21:38:43] zuul-cloner would magically fetch the patches or checkout the branche for you [21:38:46] (in theory) [21:39:13] so for mediawiki, if you send a patch to REL1_24 , zuul-cloner attempts to checkout REL1_24 for all repositories (and fallback to master) [21:39:55] (03PS5) 10Hashar: Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [21:40:03] will merge and deploy that one [21:40:16] then i am off to bed [21:41:59] (03CR) 10Hashar: [C: 032] "Let's go and congrats! I will deploy the change once merged." [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [21:43:03] (03Merged) 10jenkins-bot: Run CRM phpunit tests (non-voting) [integration/config] - 10https://gerrit.wikimedia.org/r/195042 (https://phabricator.wikimedia.org/T89896) (owner: 10Awight) [21:47:26] hashar: This can wait. The dependent change thing makes sense, but I don't see how cross-repo stuff is managed. Specifically, I will have to zuul-clone each submodule of the crm/ project, which sounds like it will do the job correctly, but is slightly annoying to maintain as we add submodules. [21:48:07] yeah hold on it as you need [21:48:17] as long as you have some kind of job that test what you want, you will be fine [21:51:05] (03PS7) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [21:51:16] (03CR) 10Awight: Jenkins job builder definition for CRM job (033 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) (owner: 10Awight) [21:52:15] awight: I haqve deployed the zuul change [21:52:26] so you should have the job triggering on the repo now :) [21:53:43] Awesome, no rush on the jjb change, there's a working version deployed. [21:54:09] refined the jjb layout change until you are happy with it [21:54:14] then I guess we can just +2 it [21:54:26] meanwhile, if you want to listen to some french song https://www.youtube.com/watch?v=QAdr5j1ztNQ [21:54:45] aired on one of our channels a few minutes ago [21:56:57] lol thx. I do a lot of internet radio, feel free to recommend stations [22:07:13] awight: well kids are both awake/crying. Time for me to leave :) [22:07:27] (03PS1) 10Legoktm: Run "php-composer-test" for MassMessage extension [integration/config] - 10https://gerrit.wikimedia.org/r/195069 [22:07:39] :( o/ [22:07:45] (03CR) 10Legoktm: [C: 032] Run "php-composer-test" for MassMessage extension [integration/config] - 10https://gerrit.wikimedia.org/r/195069 (owner: 10Legoktm) [22:08:51] (03Merged) 10jenkins-bot: Run "php-composer-test" for MassMessage extension [integration/config] - 10https://gerrit.wikimedia.org/r/195069 (owner: 10Legoktm) [22:10:15] !log deployed https://gerrit.wikimedia.org/r/195069 [22:10:21] Logged the message, Master [22:29:08] PROBLEM - Free space - all mounts on deployment-bastion is CRITICAL: CRITICAL: deployment-prep.deployment-bastion.diskspace._var.byte_percentfree.value (<33.33%) [22:48:55] (03PS1) 10Legoktm: Create generic php-composer-test-{phpflavor} [integration/config] - 10https://gerrit.wikimedia.org/r/195072 [23:04:22] 10Continuous-Integration, 10Wikimedia-Fundraising-CiviCRM: Make Civi CI job run on civicrm, drupal, and vendor (DonationInterface and SmashPig) repos - https://phabricator.wikimedia.org/T91905#1098586 (10awight) [23:19:01] (03PS8) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [23:30:57] (03CR) 10Awight: "I wish I'd noticed this earlier! Anyway, we can abandon this patch now." [integration/config] - 10https://gerrit.wikimedia.org/r/166031 (owner: 10Hashar) [23:38:09] (03PS9) 10Awight: Jenkins job builder definition for CRM job [integration/config] - 10https://gerrit.wikimedia.org/r/195063 (https://phabricator.wikimedia.org/T91895) [23:38:11] (03PS1) 10Awight: CRM job can be run on submodules [integration/config] - 10https://gerrit.wikimedia.org/r/195074 (https://phabricator.wikimedia.org/T91905) [23:39:22] (03CR) 10Legoktm: [C: 032] Create generic php-composer-test-{phpflavor} [integration/config] - 10https://gerrit.wikimedia.org/r/195072 (owner: 10Legoktm) [23:39:37] 10Continuous-Integration, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, 3Fundraising Sprint Flaming Lips, 5Patch-For-Review: Write Jenkins job builder definition for CiviCRM CI job - https://phabricator.wikimedia.org/T91895#1098631 (10awight) [23:39:59] 10Continuous-Integration, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, 3Fundraising Sprint Flaming Lips, 5Patch-For-Review: Write Jenkins job builder definition for CiviCRM CI job - https://phabricator.wikimedia.org/T91895#1098353 (10awight) [23:40:34] 10Continuous-Integration, 6Release-Engineering, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, and 3 others: Run CiviCRM testing scripts during CI - https://phabricator.wikimedia.org/T89896#1048160 (10awight) [23:41:43] 10Continuous-Integration, 6Release-Engineering, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, and 2 others: Configure Jenkins to run CiviCRM builds on Fundraising CI slave instance - https://phabricator.wikimedia.org/T89895#1048130 (10awight) [23:41:44] 10Continuous-Integration, 6Release-Engineering, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, and 3 others: Run CiviCRM testing scripts during CI - https://phabricator.wikimedia.org/T89896#1098640 (10awight) 5Open>3Resolved [23:41:50] 10Continuous-Integration, 6Release-Engineering, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, and 3 others: Deploy CiviCRM integration job to WMF integration server - https://phabricator.wikimedia.org/T86374#1098644 (10awight) [23:41:51] 10Continuous-Integration, 6Release-Engineering, 10Fundraising Tech Backlog, 10Wikimedia-Fundraising-CiviCRM, and 2 others: Configure Jenkins to run CiviCRM builds on Fundraising CI slave instance - https://phabricator.wikimedia.org/T89895#1098642 (10awight) 5Open>3Resolved a:3awight [23:42:01] eh [23:42:04] 23:39:29 Job php-composer-test-hhvm not defined [23:42:04] 23:39:29 Job php-composer-test-zend not defined [23:44:41] (03CR) 10jenkins-bot: [V: 04-1] Create generic php-composer-test-{phpflavor} [integration/config] - 10https://gerrit.wikimedia.org/r/195072 (owner: 10Legoktm) [23:45:04] (03CR) 10Legoktm: "recheck" [integration/config] - 10https://gerrit.wikimedia.org/r/195072 (owner: 10Legoktm) [23:45:36] (03CR) 10Legoktm: [C: 032] Create generic php-composer-test-{phpflavor} [integration/config] - 10https://gerrit.wikimedia.org/r/195072 (owner: 10Legoktm) [23:51:04] (03Merged) 10jenkins-bot: Create generic php-composer-test-{phpflavor} [integration/config] - 10https://gerrit.wikimedia.org/r/195072 (owner: 10Legoktm)