[00:19:42] (03CR) 10Dereckson: [C: 031] "Zoran is trying to help to the configuration and is a legitimate new contributor." [integration/config] - 10https://gerrit.wikimedia.org/r/380989 (owner: 10EddieGP) [01:33:19] PROBLEM - Puppet staleness on deployment-kafka-jumbo-1 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [43200.0] [01:53:06] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.31.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T172806#3510067 (10matmarex) Due to an issue with the patch for T48947, currently the personal links and page tabs on Hebrew WIkipedia and all RTL non-Wikipedia projects a... [01:53:12] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.31.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T172806#3641626 (10matmarex) [01:56:14] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.31.0-wmf.6 deployment blockers - https://phabricator.wikimedia.org/T174362#3559343 (10matmarex) (There is a typo in the task description here: "beneith" should be "beneath". Unfortunately I don't seem to have the rights to fix it.) [03:01:57] 10MediaWiki-Codesniffer: Mediwiki CodeSniffer MediaWiki.Classes.UnusedUseStatement.UnusedUse does not recognize Annotations - https://phabricator.wikimedia.org/T176885#3640240 (10Legoktm) @dbarratt For prioritization purposes, is this affecting a repository you're working on with MediaWiki-CodeSniffer or is this... [03:37:44] PROBLEM - English Wikipedia Mobile Main page on beta-cluster is CRITICAL: CRITICAL - Socket timeout after 10 seconds [03:42:34] RECOVERY - English Wikipedia Mobile Main page on beta-cluster is OK: HTTP OK: HTTP/1.1 200 OK - 35333 bytes in 1.338 second response time [05:14:08] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.31.0-wmf.6 deployment blockers - https://phabricator.wikimedia.org/T174362#3641734 (10mmodell) @matmarex fixed in {rPHEX162f4e7} [05:14:13] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.31.0-wmf.6 deployment blockers - https://phabricator.wikimedia.org/T174362#3641735 (10mmodell) [05:42:48] PROBLEM - Mediawiki Error Rate on graphite-labs is CRITICAL: CRITICAL: 40.00% of data above the critical threshold [10.0] [05:50:00] PROBLEM - Free space - all mounts on deployment-fluorine02 is CRITICAL: CRITICAL: deployment-prep.deployment-fluorine02.diskspace._srv.byte_percentfree (<50.00%) [06:07:49] PROBLEM - Mediawiki Error Rate on graphite-labs is CRITICAL: CRITICAL: 40.00% of data above the critical threshold [10.0] [06:32:52] PROBLEM - Mediawiki Error Rate on graphite-labs is CRITICAL: CRITICAL: 40.00% of data above the critical threshold [10.0] [06:50:02] RECOVERY - Free space - all mounts on deployment-fluorine02 is OK: OK: All targets OK [07:10:08] 10Release-Engineering-Team (Watching / External), 10Wikidata: Decide what to do with Wikibase JS-only libraries regarding the build/deployment of Wikidata code - https://phabricator.wikimedia.org/T174922#3641810 (10WMDE-leszek) Just got back from some days off and regarding what @Krinkle said in his last comme... [08:17:51] RECOVERY - Mediawiki Error Rate on graphite-labs is OK: OK: Less than 1.00% above the threshold [1.0] [08:31:10] !log Removing salt configuration from integration and deployment-prep projects. Replaced by cumin. - T176314 [08:31:15] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [08:31:17] T176314: Replace salt on integration and deployment-prep projects - https://phabricator.wikimedia.org/T176314 [08:32:50] !log Migrated Hiera config from https://wikitech.wikimedia.org/wiki/Hiera:Integration to Horizon [08:32:54] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [08:39:07] !log Deleted integration-saltmaster and deployment-salt02 . Replaced by integration-cumin and deployment-cumin - T176314 [08:39:10] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [08:39:11] T176314: Replace salt on integration and deployment-prep projects - https://phabricator.wikimedia.org/T176314 [08:41:41] PROBLEM - Host integration-saltmaster is DOWN: CRITICAL - Host Unreachable (10.68.17.68) [08:41:56] 10Beta-Cluster-Infrastructure, 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (Kanban), 10Operations-Software-Development, and 2 others: Replace salt on integration and deployment-prep projects - https://phabricator.wikimedia.org/T176314#3641943 (10hashar) @elukey has run puppet on `de... [08:43:07] PROBLEM - Host deployment-salt02 is DOWN: CRITICAL - Host Unreachable (10.68.17.58) [08:44:43] PROBLEM - Puppet errors on integration-slave-jessie-android is CRITICAL: CRITICAL: 40.00% of data above the critical threshold [0.0] [08:48:16] RECOVERY - Puppet staleness on deployment-kafka-jumbo-1 is OK: OK: Less than 1.00% above the threshold [3600.0] [08:49:20] PROBLEM - Puppet errors on deployment-mediawiki05 is CRITICAL: CRITICAL: 33.33% of data above the critical threshold [0.0] [08:50:49] (03CR) 10Hashar: [C: 032] fab: migrate from salt to cumin [integration/config] - 10https://gerrit.wikimedia.org/r/381129 (https://phabricator.wikimedia.org/T176314) (owner: 10Hashar) [08:50:52] (03CR) 10Hashar: [C: 032] fab: skip hosts when deploying slave scripts [integration/config] - 10https://gerrit.wikimedia.org/r/381131 (owner: 10Hashar) [08:52:33] (03Merged) 10jenkins-bot: fab: migrate from salt to cumin [integration/config] - 10https://gerrit.wikimedia.org/r/381129 (https://phabricator.wikimedia.org/T176314) (owner: 10Hashar) [08:52:35] (03Merged) 10jenkins-bot: fab: skip hosts when deploying slave scripts [integration/config] - 10https://gerrit.wikimedia.org/r/381131 (owner: 10Hashar) [08:55:16] 10Beta-Cluster-Infrastructure, 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (Kanban), 10Operations-Software-Development, and 2 others: Replace salt on integration and deployment-prep projects - https://phabricator.wikimedia.org/T176314#3641966 (10hashar) 05Open>03Resolved Salt ha... [09:14:06] PROBLEM - Puppet errors on deployment-mathoid is CRITICAL: CRITICAL: 44.44% of data above the critical threshold [0.0] [09:20:42] !log upgraded mediawiki04-mediawiki06 in deployment-prep to HHVM 3.18.5 [09:20:45] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [09:23:00] PROBLEM - Puppet errors on deployment-mediawiki07 is CRITICAL: CRITICAL: 60.00% of data above the critical threshold [0.0] [09:24:44] RECOVERY - Puppet errors on integration-slave-jessie-android is OK: OK: Less than 1.00% above the threshold [0.0] [09:29:21] RECOVERY - Puppet errors on deployment-mediawiki05 is OK: OK: Less than 1.00% above the threshold [0.0] [09:37:18] PROBLEM - Puppet errors on integration-slave-docker-1001 is CRITICAL: CRITICAL: 44.44% of data above the critical threshold [0.0] [09:38:28] PROBLEM - Puppet errors on deployment-pdf01 is CRITICAL: CRITICAL: 55.56% of data above the critical threshold [0.0] [09:49:05] RECOVERY - Puppet errors on deployment-mathoid is OK: OK: Less than 1.00% above the threshold [0.0] [09:57:58] RECOVERY - Puppet errors on deployment-mediawiki07 is OK: OK: Less than 1.00% above the threshold [0.0] [10:12:56] 10Release-Engineering-Team, 10Wikidata, 10User-Addshore: Move config & loading logic out of Wikidata build and into mediawiki-config - https://phabricator.wikimedia.org/T176948#3642177 (10Addshore) [10:12:58] (03CR) 10Zoranzoki21: "Thank you Dereckson very much! I will remove my review, because I can not review this request because is for me" [integration/config] - 10https://gerrit.wikimedia.org/r/380989 (owner: 10EddieGP) [10:13:26] RECOVERY - Puppet errors on deployment-pdf01 is OK: OK: Less than 1.00% above the threshold [0.0] [10:14:00] 10Release-Engineering-Team, 10Wikidata, 10Epic, 10User-Addshore: [Epic] Kill the Wikidata build step - https://phabricator.wikimedia.org/T173818#3540725 (10Addshore) [10:55:07] 10Release-Engineering-Team (Kanban), 10Readers-Web-Backlog, 10RelatedArticles, 10Browser-Tests, and 2 others: Automated browser tests cannot create pages on the Beta Cluster as anonymous user in RelatedArticles tests - https://phabricator.wikimedia.org/T176315#3642305 (10zeljkofilipin) a:03zeljkofilipin [10:55:28] 10Release-Engineering-Team (Kanban), 10Readers-Web-Backlog, 10RelatedArticles, 10Browser-Tests, and 3 others: Automated browser tests cannot create pages on the Beta Cluster as anonymous user in RelatedArticles tests - https://phabricator.wikimedia.org/T176315#3642307 (10zeljkofilipin) [11:04:35] 10Release-Engineering-Team (Kanban), 10Discovery, 10Discovery-Search (Current work), 10User-zeljkofilipin: [Epic] Port Selenium tests from Ruby to Node.js for the Search Platform - https://phabricator.wikimedia.org/T174103#3642312 (10zeljkofilipin) a:03zeljkofilipin [11:09:08] (03PS4) 10Hashar: docker: always set XDG_CACHE_HOME=/cache [integration/config] - 10https://gerrit.wikimedia.org/r/380961 [11:09:10] (03PS7) 10Hashar: Experimental integration-config-tox-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/380929 [11:09:12] (03PS5) 10Hashar: docker: stop passing HOME [integration/config] - 10https://gerrit.wikimedia.org/r/380962 [11:09:14] (03PS2) 10Hashar: docker: example with a cache volume [integration/config] - 10https://gerrit.wikimedia.org/r/380980 [11:14:51] (03Abandoned) 10Zfilipin: WIP Problem: Can not use --retry option to retry failed tests as part of the same run [selenium] - 10https://gerrit.wikimedia.org/r/341523 (https://phabricator.wikimedia.org/T160086) (owner: 10Zfilipin) [11:18:04] (03Abandoned) 10Zfilipin: WIP Headless Chrome [selenium] - 10https://gerrit.wikimedia.org/r/358578 (https://phabricator.wikimedia.org/T167507) (owner: 10Zfilipin) [11:22:46] (03PS8) 10Hashar: Experimental integration-config-tox-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/380929 [11:22:48] (03PS6) 10Hashar: docker: stop passing HOME [integration/config] - 10https://gerrit.wikimedia.org/r/380962 [11:22:50] (03PS3) 10Hashar: docker: example with a cache volume [integration/config] - 10https://gerrit.wikimedia.org/r/380980 [11:30:40] (03CR) 10Hashar: [C: 032] docker: always set XDG_CACHE_HOME=/cache [integration/config] - 10https://gerrit.wikimedia.org/r/380961 (owner: 10Hashar) [11:31:47] (03Merged) 10jenkins-bot: docker: always set XDG_CACHE_HOME=/cache [integration/config] - 10https://gerrit.wikimedia.org/r/380961 (owner: 10Hashar) [11:34:04] !log docker push wmfreleng/ci-jessie:v2017.09.28.11.33 . Adds XDG_CACHE_HOME=/cache https://gerrit.wikimedia.org/r/#/c/380961/ [11:34:08] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [11:35:20] !log docker push wmfreleng/tox:v2017.09.28.11.34 Adds XDG_CACHE_HOME=/cache https://gerrit.wikimedia.org/r/#/c/380961/ [11:35:24] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [11:36:23] (03CR) 10Hashar: "Fixed up the docker build. At least locally." [integration/config] - 10https://gerrit.wikimedia.org/r/380929 (owner: 10Hashar) [11:37:31] (03CR) 10Hashar: [C: 032] Experimental integration-config-tox-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/380929 (owner: 10Hashar) [11:40:34] (03Merged) 10jenkins-bot: Experimental integration-config-tox-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/380929 (owner: 10Hashar) [11:40:56] (03CR) 10Hashar: "check experimental" [integration/config] - 10https://gerrit.wikimedia.org/r/380929 (owner: 10Hashar) [11:42:15] (03CR) 10jenkins-bot: Experimental integration-config-tox-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/380929 (owner: 10Hashar) [11:47:19] RECOVERY - Puppet errors on integration-slave-docker-1001 is OK: OK: Less than 1.00% above the threshold [0.0] [11:54:03] 10Release-Engineering-Team, 10Wikidata, 10Patch-For-Review, 10User-Addshore: Move config & loading logic out of Wikidata build and into mediawiki-config - https://phabricator.wikimedia.org/T176948#3642397 (10Addshore) a:03Addshore [12:00:53] 10Release-Engineering-Team (Kanban), 10Discovery, 10Discovery-Search (Current work): Migrating browser tests from Jenkins on beta - https://phabricator.wikimedia.org/T175179#3642405 (10zeljkofilipin) a:05dcausse>03zeljkofilipin [12:01:33] 10Release-Engineering-Team (Kanban), 10Discovery, 10Discovery-Search (Current work), 10User-zeljkofilipin: Migrating browser tests from Jenkins on beta - https://phabricator.wikimedia.org/T175179#3642409 (10zeljkofilipin) [12:12:16] (03PS4) 10Addshore: Add all wikidata build exts to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/378771 (https://phabricator.wikimedia.org/T173940) [12:12:40] (03CR) 10Addshore: [C: 032] Add all wikidata build exts to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/378771 (https://phabricator.wikimedia.org/T173940) (owner: 10Addshore) [12:12:56] 10Release-Engineering-Team, 10Wikidata, 10Epic, 10User-Addshore: [Epic] Kill the Wikidata build step - https://phabricator.wikimedia.org/T173818#3642464 (10Addshore) [12:13:15] 10Release-Engineering-Team, 10Wikidata, 10Epic, 10User-Addshore: [Epic] Kill the Wikidata build step - https://phabricator.wikimedia.org/T173818#3540725 (10Addshore) [12:13:19] (03Merged) 10jenkins-bot: Add all wikidata build exts to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/378771 (https://phabricator.wikimedia.org/T173940) (owner: 10Addshore) [12:13:47] 10Release-Engineering-Team (Watching / External), 10Wikidata: Decide what to do with Wikibase JS-only libraries regarding the build/deployment of Wikidata code - https://phabricator.wikimedia.org/T174922#3642468 (10Addshore) a:03WMDE-leszek [12:34:19] (03CR) 10Zfilipin: Add Jenkins job for browser tests (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/375377 (https://phabricator.wikimedia.org/T160238) (owner: 10Jdlrobson) [13:03:39] (03CR) 10Zfilipin: "I would split this into two commits:" [integration/config] - 10https://gerrit.wikimedia.org/r/375377 (https://phabricator.wikimedia.org/T160238) (owner: 10Jdlrobson) [14:55:26] (03CR) 10Jayprakash12345: [C: 031] Add Zoranzoki21 to CI whitelist [integration/config] - 10https://gerrit.wikimedia.org/r/380989 (owner: 10EddieGP) [15:01:18] (03CR) 10Hashar: [C: 032] dib: use a profile to install arcanist [integration/config] - 10https://gerrit.wikimedia.org/r/380458 (owner: 10Hashar) [15:01:24] (03CR) 10Hashar: [C: 032] dib: use a profile to install zuul [integration/config] - 10https://gerrit.wikimedia.org/r/380460 (owner: 10Hashar) [15:01:37] (03CR) 10Zoranzoki21: "Thank you @Jayprakash12345" [integration/config] - 10https://gerrit.wikimedia.org/r/380989 (owner: 10EddieGP) [15:03:25] (03CR) 10Hashar: [C: 032] Capture tox logs via a full tree traversal [integration/config] - 10https://gerrit.wikimedia.org/r/380542 (owner: 10Hashar) [15:04:54] (03Merged) 10jenkins-bot: Capture tox logs via a full tree traversal [integration/config] - 10https://gerrit.wikimedia.org/r/380542 (owner: 10Hashar) [15:07:55] hashar: Could you have a look at https://gerrit.wikimedia.org/r/#/c/380989/ (new user to CI whitelist)? [15:08:25] eddiegp: in a middle of something else sorry :/ [15:08:37] PROBLEM - puppet last run on contint2001 is CRITICAL: CRITICAL: Puppet has 1 failures. Last run 2 minutes ago with 1 failures. Failed resources (up to 3 shown): Exec[jenkins-deploy docker membership] [15:09:35] hashar: Wasn't meant as "right now" ;) [15:10:50] 10MediaWiki-Codesniffer: Mediwiki CodeSniffer MediaWiki.Classes.UnusedUseStatement.UnusedUse does not recognize Annotations - https://phabricator.wikimedia.org/T176885#3643166 (10dbarratt) >>! In T176885#3641653, @Legoktm wrote: > @dbarratt For prioritization purposes, is this affecting a repository you're worki... [15:11:13] (03PS9) 10Hashar: dib: add some spec test [integration/config] - 10https://gerrit.wikimedia.org/r/380502 [15:12:23] (03CR) 10jerkins-bot: [V: 04-1] dib: add some spec test [integration/config] - 10https://gerrit.wikimedia.org/r/380502 (owner: 10Hashar) [15:12:42] (03PS10) 10Hashar: dib: add some spec test [integration/config] - 10https://gerrit.wikimedia.org/r/380502 [15:19:09] (03PS11) 10Hashar: dib: add some spec test [integration/config] - 10https://gerrit.wikimedia.org/r/380502 [15:20:22] (03CR) 10Hashar: [C: 032] "+ SPEC_OPTS=--tty to force coloring" [integration/config] - 10https://gerrit.wikimedia.org/r/380502 (owner: 10Hashar) [15:21:17] (03PS4) 10Hashar: dib: use a profile to install arcanist [integration/config] - 10https://gerrit.wikimedia.org/r/380458 [15:21:27] (03PS12) 10Addshore: Experimental Migrate 'mediawiki-core-phpcs' job to docker [integration/config] - 10https://gerrit.wikimedia.org/r/379479 (owner: 10Legoktm) [15:21:32] (03PS4) 10Hashar: dib: use a profile to install zuul [integration/config] - 10https://gerrit.wikimedia.org/r/380460 [15:21:36] (03PS5) 10Hashar: dib: use a profile to for the worker_localhost [integration/config] - 10https://gerrit.wikimedia.org/r/380467 [15:21:45] (03CR) 10Hashar: [C: 04-1] "Missing a dependency" [integration/config] - 10https://gerrit.wikimedia.org/r/380467 (owner: 10Hashar) [15:22:22] (03CR) 10jerkins-bot: [V: 04-1] dib: use a profile to for the worker_localhost [integration/config] - 10https://gerrit.wikimedia.org/r/380467 (owner: 10Hashar) [15:23:30] (03Merged) 10jenkins-bot: dib: add some spec test [integration/config] - 10https://gerrit.wikimedia.org/r/380502 (owner: 10Hashar) [15:24:08] PROBLEM - puppet last run on contint1001 is CRITICAL: CRITICAL: Puppet has 1 failures. Last run 2 minutes ago with 1 failures. Failed resources (up to 3 shown): Exec[jenkins-deploy docker membership] [15:25:05] (03CR) 10Hashar: [C: 032] "check experimental" [integration/config] - 10https://gerrit.wikimedia.org/r/380460 (owner: 10Hashar) [15:25:20] 10Browser-Tests-Infrastructure, 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (Kanban), 10Wikidata, and 3 others: Run Wikibase daily browser tests on Jenkins - https://phabricator.wikimedia.org/T167432#3643217 (10zeljkofilipin) This is what we have at the moment (from [[ https://www.m... [15:25:45] oh my god [15:29:15] legoktm: doing the phpcs job now [15:29:38] hashar: does that magic cache dir work with composer? :P [15:29:55] Sweet [15:30:23] legoktm: will get it all merged and running tonight :) [15:30:53] legoktm: want to make a little bit of bash or php code to extract the code-sniffer version so we can just install that and nothing else? :D [15:31:28] Is it really that much of a problem to install other dependencies? [15:31:49] (03PS13) 10Addshore: Experimental Migrate 'mediawiki-core-phpcs' job to docker [integration/config] - 10https://gerrit.wikimedia.org/r/379479 (owner: 10Legoktm) [15:31:52] just seems like a little bit of a wate of time :D [15:32:02] *waste.... [15:32:15] The current phpcs job that takes 10 min is a waste of time ;) [15:32:33] true [15:32:40] But yeah I can do that with jq magic [15:32:46] tbh if i fix the cache of composer packages the install should be no problem [15:32:46] 10 mins? [15:33:03] hashar: the job got pretty slow recently [15:33:04] havent we optimized it to run solely against files changed in HEAD? I thought I had a patch for that in core [15:33:19] That got reverted, it wasn't working properly [15:33:29] :((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((( [15:33:44] It was fixed, but I don't feel comfortable reenabling it until we have test coverage [15:33:52] So I don't accidentally break it again [15:34:09] ohhh [15:34:30] that got reverted whilel I was on vacations which explains why I missed the notifications :D [15:34:41] looks like it is the codesniffer upgrade that broke it :( [15:34:43] https://phabricator.wikimedia.org/T173095 [15:34:45] Mhm [15:35:01] surely that could be redone [15:35:17] I think the idea was to run phpcs directly from 'composer test' job [15:35:34] and thus drop the terrible mediawiki-core-phpcs job [15:37:05] well, the docker phgpcs job works now :D [15:37:24] but is it faster ? :] [15:37:30] should be [15:37:39] hashar: do you know if you /cache dir plans work with composer? [15:37:57] !log docker push docker.io/wmfreleng/mediawiki-phpcs:v2017.09.28.15.28 [15:38:01] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [15:38:03] so yeah, I would rather try to have some shared cache [15:38:14] I'll go check if it works with composer now [15:38:23] looking at pip, it seems we can have the cache shared on the host more or less [15:38:34] it seems to validate the content cache against the registry checksums [15:38:40] for composer ... I have no clue. Gotta test [15:38:58] hopefully it supports XDG_CACHE_HOME out of the box [15:39:16] vendor/composer/composer/doc/06-config.md:`$XDG_CACHE_HOME/composer` on unix systems that follow the XDG Base Directory [15:39:16] vendor/composer/composer/src/Composer/Factory.php: $xdgCache = getenv('XDG_CACHE_HOME') ?: $userDir . '/.cache'; [15:39:22] (from integration/composer.git) [15:39:30] addshore: so yeah composer should support the /cache trick [15:39:50] sweet [15:39:54] though really [15:40:15] sharing a cache rw between the instances is a bit crazy [15:40:26] just have it per slave [15:40:34] yeah sorry [15:40:40] I meant, sharing it between containers [15:40:55] shouldnt be too bad IMO [15:41:03] I mean, it could always be ro? [15:41:06] *read only* [15:41:20] well then how do you populate them ? :] [15:41:44] I was looking at pip this afternoon. It seems to revalidate the disk cache, so that is a start [15:42:10] but I would not guarantee there could not be some nasty (evil!) cache corruption [15:42:56] I mean, it could be read only, and populated / warmed in a similar way to how we would have to do with new docker images to the slaves [15:43:36] also, you could do some magic in the images, (probably nasty) mount the cache as read only to /readonlycache and then copy it into /cache [15:43:51] or, only the bit you need? only pip / only composer [15:44:04] how did the castor thing work? [15:44:09] wasnt that some sort of caching? [15:47:03] oh god, we need some automatied building, trying to get that cache env var up to the composer image now, and I have to rebuild the php one first :P [15:47:04] addshore: yeah [15:47:05] meh [15:47:11] addshore: populated but solely when gate and submit passed [15:47:15] which is not ideal [15:47:24] so at least the cache content is more or less vetted by someone having CR+2 [15:47:41] then jobs would be populated with the cache of the last change that ran a job in gate-and-submit [15:47:54] so the cache is outof date for a bit until a CR+2 is done [15:48:05] for composer [15:48:45] addshore: if you base your image on ci-jessie , you would get XDG_CACHE_HOME=/cache [15:48:55] (set as an ENV) [15:49:06] yup, but the chain does ci-jessie -> php -> composer [15:49:08] *goes [15:49:33] and the ENV from ci-jessie is not applied to the grand children ? [15:49:52] not automaitcally, i have to rebuild them all :p [15:49:57] ah yeah [15:50:12] at some point we would need to build a dependency tree of the docker files [15:50:17] and rebuild the ones that are obsoletes :D [15:50:18] yup [15:50:30] but really. That is starting to become crazy [15:51:09] anyway. I am out, gotta get the kids at school [15:51:28] tata! [15:51:47] twist: https://docs.docker.com/engine/reference/builder/#label :] [15:55:08] hasharAway: the cache could always not be shared, and be per workspace? :/ xD [16:08:02] (03PS14) 10Addshore: Experimental Migrate 'mediawiki-core-phpcs' job to docker [integration/config] - 10https://gerrit.wikimedia.org/r/379479 (owner: 10Legoktm) [16:10:03] (03PS15) 10Addshore: Experimental Migrate 'mediawiki-core-phpcs' job to docker [integration/config] - 10https://gerrit.wikimedia.org/r/379479 (owner: 10Legoktm) [16:11:57] legoktm https://integration.wikimedia.org/ci/job/mediawiki-core-phpcs-docker/6/console 30 seconds of setup time [16:12:47] codesniffer itself is about 2 mins I think [16:14:03] so yeh, with the cache there is no point in restricting what is installed [16:15:00] looking at https://integration.wikimedia.org/ci/job/mediawiki-core-phpcs-jessie/buildTimeTrend thats less than half the time of the jessie job [16:18:00] eww, the output of cumin is horrible :D [16:19:07] (03CR) 10Addshore: [C: 032] Experimental Migrate 'mediawiki-core-phpcs' job to docker [integration/config] - 10https://gerrit.wikimedia.org/r/379479 (owner: 10Legoktm) [16:24:05] legoktm: I might just switch it out not and see what happens [16:24:15] (03Merged) 10jenkins-bot: Experimental Migrate 'mediawiki-core-phpcs' job to docker [integration/config] - 10https://gerrit.wikimedia.org/r/379479 (owner: 10Legoktm) [16:24:58] !log Reloading Zuul to deployhttps://gerrit.wikimedia.org/r/379479 [16:35:31] (03PS1) 10Addshore: mediawiki/core run mediawiki-core-phpcs-docker instead of jessie [integration/config] - 10https://gerrit.wikimedia.org/r/381242 [16:35:33] legoktm: ^^ [16:37:31] 10Continuous-Integration-Infrastructure (shipyard): When jenkins kills a build due to max execution time the docker containers stay running - https://phabricator.wikimedia.org/T176747#3643666 (10Addshore) docker run has the following: ``` --sig-proxy Proxy received signals to the... [16:41:52] 10Continuous-Integration-Infrastructure (shipyard), 10WMDE-Analytics-Engineering, 10Patch-For-Review, 10User-Addshore, 10WMDE-QWERTY-Sprint-2017-09-19: Have CI run lintr for analytics/wmde/WDCM R files - https://phabricator.wikimedia.org/T176194#3643687 (10Addshore) The job sometimes times out and probab... [16:55:26] 10Gerrit: Data loss in gerrit review interface due to bad design - https://phabricator.wikimedia.org/T174551#3643735 (10Paladox) PolyGerrit has just recently added a way to know if a draft comment was saved or did not save. It has a little fly out box in the bottom left corner. This will be fully fix when we u... [16:56:45] addshore: lgtm! [16:56:55] cooli [16:57:03] (03CR) 10Addshore: [C: 032] mediawiki/core run mediawiki-core-phpcs-docker instead of jessie [integration/config] - 10https://gerrit.wikimedia.org/r/381242 (owner: 10Addshore) [16:57:14] your around all day to revert if it explodes and blocks people ;) [16:57:15] (03CR) 10Legoktm: mediawiki/core run mediawiki-core-phpcs-docker instead of jessie (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/381242 (owner: 10Addshore) [16:57:30] I don't have my ssh keys with me [16:57:55] (03CR) 10Addshore: [C: 032] mediawiki/core run mediawiki-core-phpcs-docker instead of jessie (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/381242 (owner: 10Addshore) [16:57:56] 10Release-Engineering-Team (Kanban), 10Release Pipeline: Install Blubber on contint1001 - https://phabricator.wikimedia.org/T175296#3643744 (10thcipriani) >>! In T175296#3609397, @dduvall wrote: > @joe or @akosiaris, would love your review of the Debian package created in {T175609} when you get a chance. We're... [16:58:10] legoktm: meh, well, i'll just wake up if it breaks then ;) [16:58:30] (03Merged) 10jenkins-bot: mediawiki/core run mediawiki-core-phpcs-docker instead of jessie [integration/config] - 10https://gerrit.wikimedia.org/r/381242 (owner: 10Addshore) [16:58:41] .... [16:58:49] !log Reloading Zuul to deploy https://gerrit.wikimedia.org/r/381242 [16:58:53] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [16:58:57] addshore: legoktm since legoktm doesn't have his keys, and it's late for you, can we... [16:59:00] well then [16:59:18] greg-g: I'll still be around for the next 7 hours! :) [16:59:22] addshore: oh good [16:59:30] I was about to trout you [16:59:57] :D [17:01:54] I need to generate new ssh keys anyways, I might as well do that now [17:02:26] well, even that 1 job should in general speed up CI quite allot... [17:02:31] ssh-keygen -o -a 100 -t ed25519 [17:02:43] addshore: yeah, that's a big win one [17:02:45] is edwhatever the new prefered thingy? [17:02:49] yup [17:03:08] * legoktm finds core patches to review [17:03:42] legoktm: hah [17:05:19] (03PS1) 10Addshore: Remove all traces of mediawiki-core-phpcs-jessie [integration/config] - 10https://gerrit.wikimedia.org/r/381255 [17:05:31] (03CR) 10Addshore: [C: 04-1] "Just gonna leave this one here for a while" [integration/config] - 10https://gerrit.wikimedia.org/r/381255 (owner: 10Addshore) [17:06:51] addshore: what php version is it using? [17:06:59] 7 [17:07:20] > Took 2 min 36 sec on integration-slave-docker-1004 [17:07:21] <3 [17:07:23] this is awesome [17:07:25] :D [17:07:38] 2 mins execution and 30 seconds setup [17:07:43] pretty consistant [17:08:04] (03PS1) 10Addshore: lintr-docker timeout 5 mins [integration/config] - 10https://gerrit.wikimedia.org/r/381256 [17:08:08] \o/ [17:08:10] is the git clone inside of the docker image our out? [17:08:12] or* [17:08:41] (03CR) 10Addshore: [C: 032] lintr-docker timeout 5 mins [integration/config] - 10https://gerrit.wikimedia.org/r/381256 (owner: 10Addshore) [17:08:50] legoktm: its inside now [17:09:14] git is now in a ci-jessie base image too :) [17:09:17] I've been thinking about the different usecases we could support if we moved it out [17:09:21] along with a couple of other things [17:09:31] legoktm: yeh, I have been troubling over that for some time too [17:09:53] like just having people do docker run [...]/core-phpcs --volume=localcheckout:whatever to get exactly the same result as CI [17:09:58] yup! [17:10:09] I'll send an email to qa@ about it [17:10:24] so, I think we should have a git and zuul-cloner image for use by CI. They can clone the source to /src, and then the test images just mount that as a volume to /src [17:10:40] seems pretty reasonable [17:10:40] well, you have to have more complicated setup images than that [17:10:49] OK, what's next on the move to docker list? [17:10:51] but for a job you would have 1 setup image and then 1 test image [17:11:05] oh also [17:11:16] well, I have been creating some general tickets at https://phabricator.wikimedia.org/project/view/3008/ that I have noticed, and they could do with some eyes and comments [17:11:20] I was wondering what to do with the slave-scripts repo [17:11:32] "shipyard"? [17:11:38] oh, I get it [17:11:38] yep xD [17:12:02] I wonder if jenkins has a view for slowest jobs [17:12:02] good name :) [17:12:17] oooh, legoktm yeh it does [17:12:29] hashar sent me something in grafana a while ago *scrolls up for days* [17:12:48] it was either slowest or most regularly run [17:13:32] https://grafana.wikimedia.org/dashboard/db/continuous-integration?orgId=1 [17:13:51] eh https://grafana.wikimedia.org/dashboard/db/nodepool-tasks?orgId=1 [17:14:07] !nodepool is https://grafana.wikimedia.org/dashboard/db/nodepool-tasks?orgId=1 [17:14:08] Key was added [17:14:19] there, now you just have to remember to ask the bot :) [17:14:40] https://grafana.wikimedia.org/dashboard/db/zuul-top-jobs?orgId=1 [17:14:49] oh, ah [17:15:42] legoktm: well, that npm one then.... [17:15:56] npm is uh [17:15:56] or mwgate-php55lint [17:16:13] and the php5lint one, thats actually quite a high number if you combine the two [17:16:31] php55lint is blazing fast, it doesn't even use nodepool right now [17:16:43] oh, not even on nodepool, lovely [17:16:54] why is npm uh? [17:17:04] the debian package is super outdated [17:17:07] mwgate-composer-validate should be trivial? [17:17:21] that one also isn't on nodepool :p [17:17:29] meh [17:17:33] did your tox patch get merged? [17:17:58] urm, yeh, there is a tox image merged, and hashar was working on setting it up for integration-config [17:18:04] ok [17:18:06] not sure what the state is there right now [17:18:21] composer-test-package-something should be easy to do [17:18:41] and then we could just try doing mw [17:19:02] * addshore looks at composer-test-package-something [17:19:20] https://phabricator.wikimedia.org/T144961 [17:19:44] that should be easy [17:19:46] gimmie a sec :p [17:20:21] wait, this is why we should do the getting of code seperate [17:20:32] we can already use the composer image to just run composer test [17:22:53] hmm, but composer-test-mwextension for example is just a builder, and is run as part of mwext-testextension-{phpflavor}-composer-{image} which also does mw-run-phpunit-allexts... gawd, which then runs mw-run-phpunit-allexts.sh..... which runs phpunit.php [17:22:54] slightly more complex :P [17:24:48] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.31.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T172806#3643855 (10greg) Status: the blocking task is not completely done (still need the cache busting), but it is no longer blocking the train. Thanks @matmarex [17:25:21] addshore: I think the task is outdated, but there's a separate composer-package-* job that just runs "composer update && composer test" [17:26:43] hmmm [17:26:48] (03CR) 10Zoranzoki21: "I am starter.. I think now to I can not to be on CI whitelist. If I can, it is super." [integration/config] - 10https://gerrit.wikimedia.org/r/380989 (owner: 10EddieGP) [17:28:21] legoktm: I'm going to have a look back at my phan image now [17:28:28] ok :) [17:28:29] as that would speed up core more again! :) [17:28:59] (03Abandoned) 10Addshore: WIP DNM docker: alternate docker based phan job.. [integration/config] - 10https://gerrit.wikimedia.org/r/378535 (owner: 10Addshore) [17:30:36] is phan the new slowest job? [17:30:55] anecdotally I think it's still mediawiki-extensions-* [17:31:07] *looks* [17:31:54] legoktm: https://integration.wikimedia.org/ci/me/my-views/view/All/ and sort by last duration? :D [17:32:15] mwext-testextension-hhvm-composer-jessie was last 15 mins [17:32:28] mwext-testextension-php55-composer-jessie last 11 mins [17:32:45] mediawiki-core-doxygen-publish last 10 mins (but doesnt run on nodepool) ? [17:33:06] because of the rsync stuff [17:33:57] well, im going to experiment with this phan one, and try fetching the code in 1 image and running phan in another [17:34:10] and the first image should be reusable in other jobs then [17:34:16] ok [17:34:55] I'll work a bit more on composer stuff later then [17:36:24] addshore: legoktm: https://integration.wikimedia.org/ci/job/integration-config-tox-docker/ runs tox in docker [17:36:45] addshore: legoktm: it is in the experimental pipeline for now. I Wanna figure out a way to cache the packages download before generalizing [17:36:57] ok :) [17:37:20] my scope being to migrate the tox* python jobs first because I <3 python [17:37:29] hasharAway: so, for the phpcs one for now I just put the cache in the workspace, not a permanent solution, but avoids the multi read write thing we were thinking about for now [17:37:50] addshore: the packages cache right? [17:38:01] hasharAway: yeh, so for the phpcs job that is composer [17:38:16] so docker run -v "$(pwd)"/cache:/cache I guess [17:38:20] yup [17:38:29] sounds good enough for now [17:38:41] hasharAway: shall I go do that for the tox job then, or let you? :) [17:39:44] I am not sure yet :] [17:39:58] I will probably use a shared cache for all the pip based jobs [17:40:12] I gotta verify the cache is indeed revalidated on each run [17:40:23] so as to avoid a container polluting the cache [17:43:56] addshore: but yeah lets do it for tox as well [17:44:14] I can over{think,engineer} later [17:44:34] oh also for the phpcs job, maybe php has xdebug enabled, which would probably not help [17:45:25] meanwhile: composer/src/Composer/Downloader/FileDownloader.php: $this->cache->gc($config->get('cache-files-ttl'), $config->get('cache-files-maxsize')); [17:48:45] composer/src/Composer/Config.php: 'cache-ttl' => 15552000, // 6 months [17:48:46] pff [17:49:20] 10Browser-Tests-Infrastructure, 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (Kanban), 10Wikidata, and 3 others: Run Wikibase daily browser tests on Jenkins - https://phabricator.wikimedia.org/T167432#3643935 (10zeljkofilipin) I have recreated [[ https://integration.wikimedia.org/ci/... [17:50:26] hasharAway, before you run off anywhere, is there any reason the zuul-cloner command uses gerrit.wikimedia.org and not contint1001 or whatever? [18:16:07] legoktm: ^^ eww it seems pointless :D [18:16:11] (03PS1) 10Addshore: DNM mediawiki-zuul-prepare-code [integration/config] - 10https://gerrit.wikimedia.org/r/381271 [18:16:13] ^^ [18:16:39] I mean, I could do exactly the same thing without building an image, and just stringing a couple of docker run commands together [18:16:45] I'm still so on the fence about how to do this [18:20:19] I feel like we should just include the integration/jenkins (slave-scripts) repository in the base image, and keep bash stuff in there, and just call the shell scripts directly [18:20:23] instead of needing a maze of docker images [18:21:26] legoktm: yeh, maybe, or maybe not in the base image but in an image used to setting up code, that has git and zuul-cloner composer and stuff [18:22:12] but then, that image doesnt have to have any set script to run, I mean you can mount any random file you want to run and change the entrypoint / command that is run anyway... [18:27:51] more or less back [18:27:58] ohia hashar! :) [18:28:40] so we gotta clone / fetch the branches from Gerrit [18:28:43] which are up to date [18:28:49] on contint1001, that is the zuul-merger process [18:29:13] and the branch can be any merge commit of the branch + a proposed patchset [18:29:19] so that is a hmm a mess [18:29:26] okay! [18:29:31] the reason is zuul-merger (on contint1001) does something like: [18:29:47] git fetch && git checkout master && git reset --hard origin/master (tip of gerrit branch) [18:29:48] then [18:30:09] git fetch refs/changes/45/12345 && git merge FETCH_HEAD (attempt to merge the proposed patchset against tip of the gerrit branch) [18:30:25] so (master) turns out to be a random merge commit not the gerrit tip of branch [18:30:30] so three steps really: [18:30:39] aaaahh, okay [18:31:03] 1) clone from cache [18:31:03] 2) bring it up to date with gerrit (git remote prune origin && git remote update) [18:31:20] 3) grab the patchset prepared as a merge against the branch [18:35:34] * hashar grabs https://github.com/editorconfig/editorconfig-vim [18:35:40] :D [18:44:49] addshore: you should send one for mediawiki :D [18:45:08] yeh, I always thought they were underused in the wikimedia world :p [18:48:41] (03PS1) 10Hashar: Convert tabs to 4 spaces [integration/config] - 10https://gerrit.wikimedia.org/r/381286 [18:49:32] addshore: well you could mentions it on wikitech-l and I am sure people will start adopting it [18:49:46] (03CR) 10Hashar: "I did not bother updating the Jenkins jobs" [integration/config] - 10https://gerrit.wikimedia.org/r/381286 (owner: 10Hashar) [18:50:59] (03PS1) 10Hashar: docker: change tox example to use master [integration/config] - 10https://gerrit.wikimedia.org/r/381287 [18:56:59] 10Browser-Tests-Infrastructure, 10Continuous-Integration-Infrastructure, 10User-zeljkofilipin: Upgrade to Chromium 59 or newer on Debian Jessie in CI - https://phabricator.wikimedia.org/T170032#3644222 (10Krinkle) p:05Normal>03High Chromium is designed as an evergreen browser. So even a one month-old ver... [18:57:09] (03PS4) 10Hashar: docker: example with a cache volume [integration/config] - 10https://gerrit.wikimedia.org/r/380980 [18:57:11] (03PS1) 10Hashar: docker ignore 'cache' and 'log' files [integration/config] - 10https://gerrit.wikimedia.org/r/381289 [18:58:12] (03CR) 10Hashar: "I haven't tested it :]" [integration/config] - 10https://gerrit.wikimedia.org/r/381289 (owner: 10Hashar) [19:02:23] (03PS2) 10Addshore: DNM mwext-php70-phan-jessie-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/381271 [19:02:30] hashar: legoktm some further thoughts ^^ [19:03:52] the creation of the bash scripts for the steps could be builders in jenkins, the scripts are then run in containers that do the setup of the source, and then the phan image runs the phan stuff :D [19:05:49] or you know [19:05:58] lets use ansible! :D [19:06:22] hashar: I can't tell if that is sarcasm or not! [19:06:35] half/half :] [19:06:50] I dont have time to dig in how Zuul digest JJB yaml dsl to build an ansible playbook [19:06:55] * thcipriani dies a little inside [19:07:25] but in theory, zuul should be able to consume all the jjb DSL and run the resulting playbook in a docker container [19:07:41] dear god [19:07:45] but I am watching at release engineering health metrics, and surely ansible is hmm sensible [19:07:45] i mean, dear goat [19:07:57] * hashar [X] learned swearing in English [19:09:36] lol [19:23:58] lol [19:36:12] 10Continuous-Integration-Config, 10Wikipedia-Android-App-Backlog: Limit Android CI jobs to running only when relevant files change - https://phabricator.wikimedia.org/T177016#3644445 (10Mholloway) [19:37:24] 10Continuous-Integration-Config, 10Wikipedia-Android-App-Backlog: Limit Android CI jobs to running only when relevant files change - https://phabricator.wikimedia.org/T177016#3644468 (10Mholloway) [19:37:40] 10Continuous-Integration-Config, 10Wikipedia-Android-App-Backlog: Limit Android CI jobs to running only when relevant files change - https://phabricator.wikimedia.org/T177016#3644445 (10Mholloway) [20:05:33] PROBLEM - Puppet errors on integration-r-lang-01 is CRITICAL: CRITICAL: 60.00% of data above the critical threshold [0.0] [20:13:45] 10Gerrit, 10Operations: Upload gerrit package to stretch apt.wm.org repo - https://phabricator.wikimedia.org/T165620#3644590 (10Dzahn) [20:13:49] 10Gerrit, 10Release-Engineering-Team (Backlog), 10Operations, 10Patch-For-Review: Reimage gerrit2001 as stretch - https://phabricator.wikimedia.org/T168562#3644589 (10Dzahn) 05Open>03stalled [20:14:05] 10Gerrit, 10Release-Engineering-Team (Backlog), 10Operations, 10Patch-For-Review: Reimage gerrit2001 as stretch - https://phabricator.wikimedia.org/T168562#3368352 (10Dzahn) Stalled by firewall on DB. [20:14:24] 10Gerrit, 10Release-Engineering-Team (Backlog), 10Operations, 10Patch-For-Review: Reimage gerrit2001 as stretch - https://phabricator.wikimedia.org/T168562#3644595 (10Dzahn) [20:14:26] 10Gerrit, 10Release-Engineering-Team (Next), 10Operations, 10Patch-For-Review: Gerrit is failing to start gerrit-ssh on gerrit2001 - https://phabricator.wikimedia.org/T176532#3644593 (10Dzahn) 05Open>03stalled stalled by firewall on DB [20:18:41] 10Gerrit, 10Release-Engineering-Team (Next), 10DBA, 10Operations, 10Patch-For-Review: Gerrit is failing to start gerrit-ssh on gerrit2001 - https://phabricator.wikimedia.org/T176532#3644612 (10Paladox) Adding DBA as we need the firewall to allow connection from m2-master.codfw.wmnet. Since the eqiad has... [20:22:50] 10Gerrit, 10Release-Engineering-Team (Next), 10DBA, 10Operations, 10Patch-For-Review: Gerrit is failing to connect to db on gerrit2001 thus preventing systemd from working - https://phabricator.wikimedia.org/T176532#3644618 (10Paladox) [20:29:13] (03PS1) 10Hashar: Switch to wmfreleng/ci-jessie [integration/quibble] - 10https://gerrit.wikimedia.org/r/381312 [20:35:33] 10Continuous-Integration-Config, 10Incident-20150312-whitespace, 10Wikimedia-Incident: add a check for whitespace before leading RECOVERY - Puppet errors on integration-r-lang-01 is OK: OK: Less than 1.00% above the threshold [0.0] [20:42:54] Project selenium-Echo » chrome,beta,Linux,BrowserTests build #531: 04FAILURE in 1 min 52 sec: https://integration.wikimedia.org/ci/job/selenium-Echo/BROWSER=chrome,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=Linux,label=BrowserTests/531/ [20:42:56] Project selenium-Echo » firefox,beta,Linux,BrowserTests build #531: 04FAILURE in 1 min 56 sec: https://integration.wikimedia.org/ci/job/selenium-Echo/BROWSER=firefox,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=Linux,label=BrowserTests/531/ [20:45:28] (03PS2) 10Hashar: Switch to wmfreleng/ci-jessie [integration/quibble] - 10https://gerrit.wikimedia.org/r/381312 [20:53:52] 10Continuous-Integration-Config, 10Incident-20150312-whitespace, 10User-greg, 10Wikimedia-Incident: add a check for whitespace before leading 03Resolved a:03greg Thanks @Seb35 I think we're good here (afaict)! [21:02:44] (03PS1) 10Hashar: Update run instruction to benefit from caching [integration/quibble] - 10https://gerrit.wikimedia.org/r/381352 [21:25:51] thcipriani: you around? [21:26:05] tabbycat: sure, what's up? [21:26:22] thcipriani: wondering if you could hit the 'rebase' button in https://gerrit.wikimedia.org/r/#/c/292779/ please? [21:27:05] since it's not my change, the button does not appear for me [21:27:10] done :) [21:27:22] thanks :D [21:27:32] sure [21:30:03] 10Gerrit, 10Upstream: Gerrit should track follow-ups to a merged change (by commit SHA-1 and Change-Id), like CodeReview used to - https://phabricator.wikimedia.org/T64371#3644817 (10Paladox) Filled upstream at https://bugs.chromium.org/p/gerrit/issues/detail?id=7299 against polygerrit. GWTUI is in the proces... [21:30:49] we should grant rebasing to everyone as you can easily hit the cherry pick and click the same branch [21:31:03] which basically updates the parent but not totally a rebase. [21:31:31] that's the workaround ^^ [21:34:25] * paladox submits the change here https://gerrit.wikimedia.org/r/#/c/381356/ [21:48:13] any objections if I deploy a security patch? [21:51:14] go forth and securitize [21:51:25] goat-for-it [22:15:54] (03PS3) 10Addshore: DNM mwext-php70-phan-jessie-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/381271 [22:16:15] (03CR) 10Addshore: "Looks like the docker nodes don't have enough memory to run phan?" [integration/config] - 10https://gerrit.wikimedia.org/r/381271 (owner: 10Addshore) [22:25:51] (03PS1) 10Addshore: docker: use nobody user instead of creating custom users [integration/config] - 10https://gerrit.wikimedia.org/r/381363 [22:31:09] (03PS4) 10Addshore: DNM mwext-php70-phan-jessie-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/381271 [22:31:21] (03PS5) 10Addshore: DNM mwext-php70-phan-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/381271 [22:36:22] legoktm: looks like my phan job for extensions works just fine [22:36:57] and, i think, it might be a goodish template for the way to setup mediawiki for tests [22:37:42] basically have an image that has git, composer, zuul-cloner, npm and whatever else you might need to install / setup code, and then have a separate image for running tests. [22:38:07] Sounds good to me :)) [22:38:26] It would also mean my crazy idea earlier of having an initial job setup the code and then a bunch of jobs run using that 1 set of code in read only mode might not even bet that crazy, but i wouldnt know how to tackle that... [22:39:06] also, it looks like the docker hosts only have 2gb ram, but nodepool had 4gb ram, and phan might like to have more otherwise it might die in some cases :/ [22:41:02] 10Continuous-Integration-Infrastructure (shipyard): phan can fail to allocate memory on current docker hosts - https://phabricator.wikimedia.org/T177031#3644958 (10Addshore) [22:41:39] (03CR) 10Addshore: [C: 032] "new images pushed with "nobody" being used in them" [integration/config] - 10https://gerrit.wikimedia.org/r/381363 (owner: 10Addshore) [22:42:24] (03PS2) 10Addshore: lintr-docker timeout 5 mins [integration/config] - 10https://gerrit.wikimedia.org/r/381256 [22:42:32] (03CR) 10Addshore: [C: 032] lintr-docker timeout 5 mins [integration/config] - 10https://gerrit.wikimedia.org/r/381256 (owner: 10Addshore) [22:42:44] (03Merged) 10jenkins-bot: docker: use nobody user instead of creating custom users [integration/config] - 10https://gerrit.wikimedia.org/r/381363 (owner: 10Addshore) [22:44:04] (03Abandoned) 10Addshore: docker: mediawiki-extensions-phan image [integration/config] - 10https://gerrit.wikimedia.org/r/371708 (owner: 10Addshore) [22:44:17] (03Merged) 10jenkins-bot: lintr-docker timeout 5 mins [integration/config] - 10https://gerrit.wikimedia.org/r/381256 (owner: 10Addshore) [22:59:08] (03PS6) 10Addshore: DNM mwext-php70-phan-docker job [integration/config] - 10https://gerrit.wikimedia.org/r/381271