[00:09:50] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: Vagrant 1.8.7 fails to fetch Jessie image with vague error message - https://phabricator.wikimedia.org/T158608#3041762 (10brion) [00:22:17] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: Vagrant 1.8.7 fails to fetch Jessie image with vague error message - https://phabricator.wikimedia.org/T158608#3041776 (10bd808) I don't see anything at https://atlas.hashicorp.com/debian/boxes/contrib-jessie64 that explicitly says a re... [00:51:46] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: [EPIC] Migrate base image to Debian Jessie - https://phabricator.wikimedia.org/T136429#3041805 (10brion) [00:51:50] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: Vagrant 1.8.7 fails to fetch Jessie image with vague error message - https://phabricator.wikimedia.org/T158608#3041803 (10brion) 05Open>03Invalid Running with --debug seems to indicate that Vagrant's downloader is failing to load cu... [00:59:22] bd808: for https://gerrit.wikimedia.org/r/#/c/338820 as amended, your severity patch is not necessary, is it? [00:59:34] I can just deploy it and see what we get? [01:01:22] TimStarling: yeah. give it a shot and lets see how fast the hard drive fills up :) [01:02:01] Its been a long time since Reedy and Krinkle last tried it [01:02:05] heh [01:02:22] I think we had a few gig in 20-30 minutes [01:02:33] TimStarling: Might be worth doing a few X-Wikimedia-Debug requests with logging enabled to get an idea of how much [01:02:40] afaik they are already reported to logstash in that case [01:02:43] that's how I found T158105 for example [01:02:44] T158105: "PHP Warning: filemtime(): No such file or directory" about files removed over a year ago - https://phabricator.wikimedia.org/T158105 [01:03:21] We intentionally surpress that filemtime warning and instead do a === false comparison in the code (per the Python-inspired pattern of try first instead of checking and then doing since race conditions and performance) [01:04:48] but yeah, coudl also try on entire prod and be ready with a revert :) [01:05:57] actually this change will (hopefully) not log suppressed errors [01:06:19] because suppressed errors only appear in the error-json channel, not plain "error" [01:08:43] Yeah [01:08:50] interesting to know that you can find bugs that way though [01:09:01] I was making fun of the idea in this channel earlier [01:09:14] TimStarling: Sorry, that was confusing. I meant, you sync-pull the patch from tin to mwdebug1001 and run a few XWD requests first to see what ends up logged. [01:10:22] TimStarling: Yeah, I think it's a rare case and probably not worth logging on all of prod. But yeah, I think there are certainly cases where the author of code intends to handle an error and has reason to surpress the warning, but routine inspection can still help find warnings that are clearly unexpected. [01:11:05] In my case, the code is expected to find a non-existent path every once in a while given there is a few minutes of caching in between. But if I see it logging a file path I know has been removed >6 months ago, that's a problem with the cache. [01:11:37] But we already make that easy by logging everything (incl.. suppressed) with XWMD set [01:12:05] or rather, if 'log' is set in the XWMD header [01:12:21] if you enable XWMD but don't set the log checkbox, you'll get the normal log flow. [01:13:08] I'm planning on pushing that change I just merged to all servers on the assumption that what will be logged will be a subset of what is already going to fluorine in hhvm.log [01:13:55] right. though larger since it includes more per event (stack traces) [01:14:53] currently hhvm.log is 0.02% of the total [01:18:32] works [01:18:56] current flow rate would not overload my raspberry pi based syslog server [01:25:43] TimStarling: thanks for being bold and giving it another try :) [01:28:09] * Reedy wonders what fun stuff this will uncover [01:29:38] Redis and Memcached stuff it looks like [01:58:39] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: Job runner service doesn't appear to work in jessie-migration - https://phabricator.wikimedia.org/T158615#3041900 (10brion) [02:03:29] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: Job runner service doesn't appear to work in jessie-migration - https://phabricator.wikimedia.org/T158615#3041913 (10brion) Note there is no `logs/mediawiki-runJobs.log` file, and I cannot connect to port 80 on 127.0.0.1 from within the... [02:04:26] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 50.00% of data above the critical threshold [0.0] [02:29:09] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: Job runner service doesn't appear to work in jessie-migration - https://phabricator.wikimedia.org/T158615#3041919 (10bd808) Likely broken by {rMWVA1956f986abfe} where we dropped the port 80 bind. [03:09:27] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0] [03:52:44] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations, 07Epic: [EPIC] Migrate base image to Debian Jessie - https://phabricator.wikimedia.org/T136429#3041958 (10brion) [04:07:12] Project selenium-MultimediaViewer » safari,beta,OS X 10.9,BrowserTests build #307: 04FAILURE in 11 min: https://integration.wikimedia.org/ci/job/selenium-MultimediaViewer/BROWSER=safari,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=OS%20X%2010.9,label=BrowserTests/307/ [04:33:49] 06Release-Engineering-Team, 10MediaWiki-Vagrant, 06Operations: npm install fails for changeprop service in MW-Vagrant jessie-migration - https://phabricator.wikimedia.org/T158617#3041968 (10brion) [06:30:27] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 20.00% of data above the critical threshold [0.0] [06:30:31] PROBLEM - Free space - all mounts on deployment-fluorine02 is CRITICAL: CRITICAL: deployment-prep.deployment-fluorine02.diskspace._srv.byte_percentfree (<50.00%) [06:39:04] 10Continuous-Integration-Config, 10ContentTranslation, 03Language-2017 Sprint 2, 03Language-2017 Sprint 3, and 4 others: mwext-qunit-jessie test fails on unrelated change - https://phabricator.wikimedia.org/T153038#3042066 (10Arrbee) [06:50:33] RECOVERY - Free space - all mounts on deployment-fluorine02 is OK: OK: All targets OK [07:10:27] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0] [08:08:41] Project selenium-MultimediaViewer » firefox,mediawiki,Linux,BrowserTests build #308: 04FAILURE in 2 min 6 sec: https://integration.wikimedia.org/ci/job/selenium-MultimediaViewer/BROWSER=firefox,MEDIAWIKI_ENVIRONMENT=mediawiki,PLATFORM=Linux,label=BrowserTests/308/ [08:18:42] Yippee, build fixed! [08:18:43] Project selenium-MultimediaViewer » safari,beta,OS X 10.9,BrowserTests build #308: 09FIXED in 12 min: https://integration.wikimedia.org/ci/job/selenium-MultimediaViewer/BROWSER=safari,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=OS%20X%2010.9,label=BrowserTests/308/ [08:36:30] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 66.67% of data above the critical threshold [0.0] [08:42:53] zeljkof: migration is complete. What is left are the selenium-*-TXXXX jobs [08:43:03] which are still on the trusty slave [08:43:05] hashar: those are all testing jobs [08:43:09] but there is no point in migrating them imho [08:43:23] then on https://integration.wikimedia.org/ci/view/Selenium/ [08:43:26] I can delete them and recreate on new infra, or just delete when done [08:43:54] I made it to filter out slaves that are not running selenium-* jobs :} so at one point the build executors list on the left would only shows the saucelabs-XX instances [08:44:09] zeljkof: just delete them once done [08:44:20] there is no point wasting time migrating them [08:44:30] ok [08:44:35] was just mentioning I haven't migrated them :D [09:05:41] (03PS2) 10Hashar: Use puppet-syntax to check puppet manifests [integration/config] - 10https://gerrit.wikimedia.org/r/338746 (https://phabricator.wikimedia.org/T158544) [09:09:33] (03PS3) 10Hashar: Use puppet-syntax to check puppet manifests [integration/config] - 10https://gerrit.wikimedia.org/r/338746 (https://phabricator.wikimedia.org/T158544) [09:11:07] (03CR) 10Hashar: [C: 032] Use puppet-syntax to check puppet manifests [integration/config] - 10https://gerrit.wikimedia.org/r/338746 (https://phabricator.wikimedia.org/T158544) (owner: 10Hashar) [09:11:28] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0] [09:12:50] (03CR) 10Hashar: [C: 032] Use puppet-syntax to check puppet manifests [integration/config] - 10https://gerrit.wikimedia.org/r/338746 (https://phabricator.wikimedia.org/T158544) (owner: 10Hashar) [09:14:16] (03Merged) 10jenkins-bot: Use puppet-syntax to check puppet manifests [integration/config] - 10https://gerrit.wikimedia.org/r/338746 (https://phabricator.wikimedia.org/T158544) (owner: 10Hashar) [09:32:27] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 33.33% of data above the critical threshold [0.0] [09:39:31] hashar: how can i / should i request the creation of a new beta site? [09:39:36] I would quite like a hewiktionary! [09:39:54] uselang=he ? :D [09:40:00] bwhahaa, no ;) [09:40:05] to have hewiktionary settings ;) [09:40:13] specifically for testing the InteriwkiSorting & Cognate extension [09:40:27] 1) fill a task [09:40:31] 2) follow https://wikitech.wikimedia.org/wiki/Nova_Resource:Deployment-prep/Add_a_wiki [09:40:35] and you should be fine :} [09:40:37] epic [09:40:58] there are a few moving parts ( have to add swift, parsoid, restbase) [09:41:02] so it is not so easy :/ [09:42:56] hashar: what project on phab should I be adding? [09:43:06] #beta-cluster-infrastructure [09:43:11] =] [09:43:21] it is merely for history purpose [09:43:27] you could just !log it here [09:43:40] but a task gives the whole background, let you paste the commands that have been run etc [09:43:43] 10Beta-Cluster-Infrastructure, 10Cognate, 10MediaWiki-extensions-InterwikiSorting, 15User-Addshore: Create beta hewiktionary for testing InterwikiSorting & Cognate - https://phabricator.wikimedia.org/T158628#3042289 (10Addshore) [09:43:51] \\O// [09:43:56] so, in theroy I can manage to set this all up myself? [09:44:15] hmm [09:44:20] hopefully ? :-} [09:44:29] 10Beta-Cluster-Infrastructure, 10Cognate, 10MediaWiki-extensions-InterwikiSorting, 15User-Addshore: Create beta hewiktionary for testing InterwikiSorting & Cognate - https://phabricator.wikimedia.org/T158628#3042302 (10Addshore) [09:44:59] lets see :p [09:45:04] to setup swift on beta you can probably ask support from godog [09:45:11] okay [09:45:13] he is our files / swift god nowadays :D [09:45:50] for Parsoid, I have no clue how the service is deployed / updated [09:45:52] maybe it is a cron, maybe it is manually done [09:46:07] thehe the swift part should be self-service if you have access to mediawiki-config with the credentials [09:46:13] RESTBase you apparently need a patch to puppet.git , and hence a cherry pick on beta [09:46:15] iow mw can create what it needs by itself [09:46:37] godog: yeah looks like there are some mw maintenance script to take care of swift! https://wikitech.wikimedia.org/wiki/Add_a_wiki#Swift [09:47:05] addshore: and beware, the wiki doc might have outdated bits [09:47:06] yep, that exactly [09:47:12] magic! [09:47:27] okay, well, I'll start going through it and see how it goes! [09:49:16] hashar: looks like I failed at step 1 ;) I cant edit https://wikitech.wikimedia.org/wiki/Hiera:Deployment-prep/host/deployment-cache-text04 ;) [09:52:13] 10Continuous-Integration-Config, 10ContentTranslation, 03Language-2017 Sprint 2, 03Language-2017 Sprint 3, and 4 others: mwext-qunit-jessie test fails on unrelated change - https://phabricator.wikimedia.org/T153038#3042313 (10Nikerabbit) `{responseText: '', status: 0, statusText: 'error'}` again. There is... [09:53:30] 10Beta-Cluster-Infrastructure, 10Cognate, 10MediaWiki-extensions-InterwikiSorting, 15User-Addshore: Create beta hewiktionary for testing InterwikiSorting & Cognate - https://phabricator.wikimedia.org/T158628#3042314 (10Addshore) It looks like I fell at the first hurdle of the steps listed @ https://wikitec... [09:55:31] 10Continuous-Integration-Config, 10ContentTranslation, 03Language-2017 Sprint 2, 03Language-2017 Sprint 3, and 4 others: mwext-qunit-jessie test fails on unrelated change - https://phabricator.wikimedia.org/T153038#3042333 (10Nikerabbit) Testing in https://gerrit.wikimedia.org/r/#/c/338937/ – perhaps reque... [09:58:06] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure: mediawiki-extensions-qunit-jessie tests are failing - https://phabricator.wikimedia.org/T153597#3042345 (10Nikerabbit) Should this be closed then? [10:07:26] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0] [10:30:01] (03CR) 10Hashar: [C: 04-1] "Npm has been added eventually. There is a composer.json file in the repo but it does not define a composer test entry point yet." [integration/config] - 10https://gerrit.wikimedia.org/r/280139 (owner: 10Paladox) [10:46:16] (03Abandoned) 10Hashar: [FundraisingChart] Add npm test [integration/config] - 10https://gerrit.wikimedia.org/r/275574 (owner: 10Paladox) [10:50:05] (03PS11) 10Hashar: Test mediawiki gate queue only has mw projects [integration/config] - 10https://gerrit.wikimedia.org/r/241660 (https://phabricator.wikimedia.org/T107529) [10:52:08] (03CR) 10Hashar: "Good enough for now. It is harnessed behind a 'qa' attribute so would only run via 'tox -e qa'." [integration/config] - 10https://gerrit.wikimedia.org/r/241660 (https://phabricator.wikimedia.org/T107529) (owner: 10Hashar) [10:54:41] (03CR) 10Hashar: [C: 032] Test mediawiki gate queue only has mw projects [integration/config] - 10https://gerrit.wikimedia.org/r/241660 (https://phabricator.wikimedia.org/T107529) (owner: 10Hashar) [10:55:52] (03Merged) 10jenkins-bot: Test mediawiki gate queue only has mw projects [integration/config] - 10https://gerrit.wikimedia.org/r/241660 (https://phabricator.wikimedia.org/T107529) (owner: 10Hashar) [11:05:22] ./grunt-eslint/node_modules/eslint/lib/config/config-file.js: ".eslintrc.jso", [11:05:25] seriously [12:16:48] addshore: do you have any idea what is wrong with https://integration.wikimedia.org/ci/job/mediawiki-extensions-hhvm-jessie/6438/consoleFull ? [12:16:55] *looks* [12:17:03] seems not to have or find the hamcrest code [12:17:10] but composer appears to install it [12:17:34] hmmmm [12:18:05] tests work locally (i have the build setup) [12:18:18] have you tried on hvvm locally? [12:18:22] yes [12:18:25] hmm [12:18:35] nothing obvious jumps out [12:18:36] hamcrest provides global functions (meh) [12:18:41] but they should work [12:19:10] not having a passing build will prevent being able to deploy :/ [12:19:57] wait, this is for the build? [12:20:02] yeah [12:20:05] it pases in wikibase then? [12:20:10] yes [12:20:16] maybe some path is wrong [12:20:29] but hamcrest is installed into core [12:20:32] hmmm, okay, I can probably have a hunt around, how urgent is it? / when do you want to deploy? [12:20:34] so shouldn't matter [12:20:48] i'm poking also, [12:20:56] we want to deploy to test.wikidata later [12:24:32] apparently it's been failing for days (since hamcrest was introduced) :( [12:24:35] https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/37316/console [12:24:44] * aude assumed it was mediainfo still failing [12:38:10] rebuilding and will try to look at the workspace [12:41:18] missing froom https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/ws/src/vendor/composer/autoload_files.php/*view*/ [12:48:51] hmmm [12:49:41] aude: as a guess then it is something to do with it being included here https://github.com/wikimedia/mediawiki/blob/master/composer.json#L79 [12:49:50] * aude looks [12:50:28] in some cases, i assume it does get included / installed [12:51:56] aude: also as a guess it could be something vaugly related to https://github.com/composer/composer/issues/3922 [12:52:11] but perhaps not [12:52:29] possible [12:52:42] but how it works for wikibase tests [12:54:31] https://github.com/wikimedia/integration-jenkins/blob/master/bin/mw-fetch-composer-dev.sh [12:54:43] that looks like it should be okay [12:55:24] FIXME: integration/composer is outdated and breaks autoloader seems interesting [12:55:41] currently on 1.0.3 [12:55:46] oh [12:55:52] i could try that version of composer [12:55:58] yeh, that could be it [12:56:29] current version is 1.3.2 [12:56:37] o [12:56:38] k [13:05:36] * aude off to meeting [13:22:05] 10Gerrit, 06Release-Engineering-Team, 10DBA, 06Operations, 13Patch-For-Review: Gerrit shows HTTP 500 error when pasting extended unicode characters - https://phabricator.wikimedia.org/T145885#3042690 (10Paladox) We can still do https://gerrit.wikimedia.org/r/#/c/336002/ since I doint see an urgency to ha... [13:41:25] 10Browser-Tests-Infrastructure, 15User-zeljkofilipin: Write a plan on how to move Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T158644#3042727 (10zeljkofilipin) [13:41:42] 10Browser-Tests-Infrastructure, 15User-zeljkofilipin: Write a plan on how to move Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T158644#3042743 (10zeljkofilipin) p:05Triage>03High [14:38:02] what are these saucelab instances [14:57:08] 06Release-Engineering-Team, 07Jenkins: Depool precise jenkins instances - https://phabricator.wikimedia.org/T158652#3042963 (10Zppix) [15:08:53] hashar: ^^ [15:10:21] I don't think there is a need for such a task but well [15:10:25] I guess it doesn't hurt either [15:10:49] hashar: to be fair it needs to happen before labs removes precise [15:14:14] no that is the other way around [15:14:28] labs is not going to remove precise while we still use them :} [15:14:51] going to talk about it during the releng team meeting in a couple hours from now [15:15:15] there is some task about phasing out all precise instances on which we commented on the specific CI use case. I wrote a summary in some of the comments [15:15:29] could you email me and let me know what the plan is after the meeting hashar? [15:15:44] hashar: task id? [15:19:02] you will know [15:19:08] we will announce it [15:20:09] 10Gerrit, 06Release-Engineering-Team, 06Operations: Decide weather to disables drafts in gerrit - https://phabricator.wikimedia.org/T158656#3043080 (10Paladox) [15:23:40] 10Gerrit, 06Release-Engineering-Team, 06Operations: Decide weather to disables drafts in gerrit - https://phabricator.wikimedia.org/T158656#3043111 (10Paladox) Here is the changes https://gerrit-review.googlesource.com/#/q/topic:private-changes+(status:open+OR+status:merged) that will bring support for priva... [15:24:00] 10Gerrit, 06Release-Engineering-Team, 06Operations: Decide weather to disable drafts in gerrit - https://phabricator.wikimedia.org/T158656#3043112 (10Paladox) [15:27:01] 10Browser-Tests-Infrastructure, 15User-zeljkofilipin: Write a plan on how to move Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T158644#3043117 (10zeljkofilipin) 05Open>03Invalid [15:27:26] 10Continuous-Integration-Infrastructure, 06Release-Engineering-Team: Depool precise jenkins instances - https://phabricator.wikimedia.org/T158652#3043118 (10hashar) Pending outcome of T143349#3024759 . There is a lot of CI configuration work to happen before we can consider dropping the Precise instances. [15:44:32] Project selenium-MobileFrontend » chrome,beta,Linux,BrowserTests build #335: 04FAILURE in 22 min: https://integration.wikimedia.org/ci/job/selenium-MobileFrontend/BROWSER=chrome,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=Linux,label=BrowserTests/335/ [16:01:23] 10Browser-Tests-Infrastructure, 10Wikidata, 13Patch-For-Review, 15User-Tobi_WMDE_SW, 15User-zeljkofilipin: Increase in failures caused by Saucelabs - https://phabricator.wikimedia.org/T152963#3043255 (10zeljkofilipin) > Albert Sison (Sauce Labs Help Center) > > Feb 17, 3:22 PM PST > Hello, > > Thanks f... [16:21:05] 10Gerrit, 06Release-Engineering-Team, 13Patch-For-Review, 07Regression, 07Upstream: Cannot log into Gerrit as of recent upgrade - https://phabricator.wikimedia.org/T152640#3043284 (10demon) Yes, it is. [16:29:41] 10Browser-Tests-Infrastructure, 10MediaWiki-General-or-Unknown, 07JavaScript, 13Patch-For-Review, and 2 others: Port Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T139740#3043304 (10zeljkofilipin) [16:36:26] (03PS1) 10Umherirrender: [DuskToDawn] Add npm job [integration/config] - 10https://gerrit.wikimedia.org/r/338991 [16:45:57] (03CR) 10Hashar: [C: 032] [DuskToDawn] Add npm job [integration/config] - 10https://gerrit.wikimedia.org/r/338991 (owner: 10Umherirrender) [16:51:12] (03Merged) 10jenkins-bot: [DuskToDawn] Add npm job [integration/config] - 10https://gerrit.wikimedia.org/r/338991 (owner: 10Umherirrender) [17:32:42] 10Browser-Tests-Infrastructure, 10MediaWiki-General-or-Unknown, 07JavaScript, 13Patch-For-Review, and 2 others: Port Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T139740#3043638 (10zeljkofilipin) [17:39:24] 10Continuous-Integration-Config, 06Wikipedia-Android-App-Backlog, 07Technical-Debt: Figure out the right way to keep the Android SDK up to date in CI - https://phabricator.wikimedia.org/T158456#3043664 (10Niedzielski) I think we actually end up with a couple copies of the SDK on a dev machine: 1. Android... [17:40:24] 10Browser-Tests-Infrastructure, 10MediaWiki-General-or-Unknown, 07JavaScript, 13Patch-For-Review, and 2 others: Port Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T139740#3043667 (10zeljkofilipin) [17:46:09] 06Release-Engineering-Team (Deployment-Blockers), 05Release: MW-1.29.0-wmf.11 deployment blockers - https://phabricator.wikimedia.org/T155526#3043710 (10greg) 05Open>03Resolved [17:46:17] 06Release-Engineering-Team (Deployment-Blockers), 13Patch-For-Review, 05Release: MW-1.29.0-wmf.12 deployment blockers - https://phabricator.wikimedia.org/T155527#3043711 (10greg) 05Open>03Resolved [17:54:51] 10Browser-Tests-Infrastructure, 10MediaWiki-General-or-Unknown, 07JavaScript, 13Patch-For-Review, and 2 others: Port Selenium tests from Ruby to Node.js - https://phabricator.wikimedia.org/T139740#3043730 (10zeljkofilipin) [18:25:55] 10Deployment-Systems, 06Release-Engineering-Team, 10MediaWiki-Special-pages: All extension should link to the deployed Git revision from Special:Version - https://phabricator.wikimedia.org/T155747#3043918 (10Aklapper) [19:02:23] 10Continuous-Integration-Config, 06Wikipedia-Android-App-Backlog, 07Technical-Debt: Figure out the right way to keep the Android SDK up to date in CI - https://phabricator.wikimedia.org/T158456#3044040 (10Mholloway) A lot of stuff got downloaded (and licenses accepted) last Friday while I was looking at this... [19:11:38] I can't vagrant up because of a bug [19:11:48] vagrant up wants to update the guest additions because the version is old [19:12:00] Or rather, because they were upgraded on the host [19:12:01] I can't either because parallels decided to break support for vagrant for no reason :( [19:12:06] So it tries to apt-get update [19:12:17] But that fails because /vagrant isn't NFS-mounted yet [19:12:39] And I can't get it to mount because vagrant up doesn't get that far (and also presumably I might need the guest additions to perform that mount in the first place) [19:12:42] RoanKattouw: yeah. that bug sucks. The fix is to disable the guest update plugin. Let me find the paste that you need [19:12:43] Help? [19:12:49] Oh cool [19:12:49] Thanks [19:13:22] RoanKattouw: https://phabricator.wikimedia.org/P2955 [19:13:45] make a Vagrantfile-extra.rb in the same directory as your Vagrantfile with that content [19:14:56] Ah, -extra not .extra, got it [19:15:10] That'll teach me to transcribe rather than copy paste [19:15:54] Yay, that worked, thanks! [19:34:30] 10Continuous-Integration-Config, 10Continuous-Integration-Infrastructure: mediawiki-extensions-qunit-jessie tests are failing - https://phabricator.wikimedia.org/T153597#3044134 (10Tgr) 05Open>03Resolved [19:37:59] 10Continuous-Integration-Config, 06Wikipedia-Android-App-Backlog, 07Technical-Debt: Figure out the right way to keep the Android SDK up to date in CI - https://phabricator.wikimedia.org/T158456#3044145 (10Niedzielski) Maybe the emulator used is intentionally outdated so that it will be compatible with the Je... [20:15:05] PROBLEM - Puppet run on integration-slave-trusty-1003 is CRITICAL: CRITICAL: 50.00% of data above the critical threshold [0.0] [20:28:47] bah [20:40:04] RECOVERY - Puppet run on integration-slave-trusty-1003 is OK: OK: Less than 1.00% above the threshold [0.0] [20:45:04] RainbowSprinkles: greg-g: Would like a window after today and thursday's train for https://phabricator.wikimedia.org/T158105#3037743 [20:45:13] won't take but a few minutes [20:45:21] maintenance script for wmf.13 resourceloader [20:45:34] Today's train is done, go right ahead [20:45:41] Okay. [21:04:28] RECOVERY - Long lived cherry-picks on puppetmaster on deployment-puppetmaster02 is OK: OK: Less than 100.00% above the threshold [0.0] [21:08:11] bd808: maybe you or someone can help me with a composer / jenkins issue? [21:08:37] aude: Maybe you'll have more luck with oil and water [21:08:42] aude: chances are slim, but what's the problem? [21:08:57] https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/37857/console is failing because hamcrest is not being autoloaded into mediawiki's vendor [21:09:23] when the wikidata build tests are run, but it works otherwise when other stuff is run (e.g. wikibase directly ) [21:09:30] https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/ws/src/vendor/composer/autoload_files.php/*view*/? [21:09:48] i know we use mediawiki/vendor and then add mediawiki's dev dependencies [21:09:57] https://github.com/wikimedia/integration-jenkins/blob/master/bin/mw-fetch-composer-dev.sh [21:10:24] works locally, i tried with same version of composer and i have hhvm [21:20:26] just wonder if parts of the workspace are left over from a previous job or are cached [21:22:35] aude: on that job, workspace is left behind [21:22:42] and is reused between builds [21:23:05] ok [21:24:48] aude: when building the ham crest patches, one of the issue was that the lib was a require-dev in Wikibase [21:25:05] and the composer merge plugin is configured in mediawiki/core to NOT merge dev dependencies [21:25:08] yeah [21:25:17] so I proposed to add that ham crest lib directly in mediawiki/core [21:25:34] yeah, it's part of core now and can be used to replace assertTag there [21:25:35] on two basis: a) it will eventually be used by mediawiki/core itself to migrate the PHPUnit assertTag() [21:25:40] yeah [21:25:41] b) can be reused by other exts [21:25:50] as for mediawiki/vendor I have no clue :( [21:26:05] mediawiki/vendor is only non-dev things afaik [21:26:16] that is my understanding as well [21:26:30] basically a result of a composer merge plugin run of core + whatever we got deployed on wmf [21:26:33] plus the script that adds dev things for jenkins [21:26:52] also, per wikitech mail from Krinkle from a year ago [21:27:07] and that Fatal error: Call to undefined function Wikibase\Client\Tests\RecentChanges\both() ? [21:27:21] is that really due to lack of hamcrest? [21:27:32] https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/37857/console says hamcrest and wmde/hamcrest... are installed [21:27:48] loading from cache [21:28:19] i retriggered the build and checked the workspace and it was missing hamcrest [21:28:29] eeek [21:29:29] https://github.com/hamcrest/hamcrest-php/blob/master/hamcrest/Hamcrest.php#L241-L252 [21:29:34] would capturing the composer.lock help ? [21:29:40] maybe [21:30:01] composer autoload files looked like https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/ws/src/vendor/composer/autoload_files.php/*view*/ [21:30:19] should have 2 additional entries there for hamcrest [21:33:37] https://integration.wikimedia.org/ci/job/mwext-testextension-hhvm/ws/src/vendor/composer.lock/*view*/ [21:40:25] 10Deployment-Systems, 10Scap, 10scap2: scap wikiversions compile happening too late in scap sync - https://phabricator.wikimedia.org/T156851#3044701 (10thcipriani) [21:43:00] eek I got a change but it is rather invasive :( [21:43:42] :/ [21:43:42] hashar: has the releng meeting happened if so can i have the result if possible? [21:44:02] Zppix he wrote on a different task. [21:44:08] * paladox goes and finds it [21:44:26] Ok [21:44:27] Zppix https://phabricator.wikimedia.org/T143349#3043382 [21:46:01] Would the php test ran on tools.zppix bot repo change (under labs tab on jenkins on integration.wikimedia.org [21:46:33] Zppix not sure what you mean? [21:46:59] Labs-tools-ZppixBot-php55lint [21:47:01] (03PS1) 10Hashar: Capture composer.lock after mw-fetch-composer-dev [integration/config] - 10https://gerrit.wikimedia.org/r/339051 [21:47:07] No, cause you're not using a php 5.3 lint [21:47:14] Ok [21:47:21] Zppix php55lint will be un affected [21:47:28] that is run on trusty [21:47:29] Zppix: php 5.3 support will be dropped [21:47:30] Thats what i thought [21:47:39] 53 runs on precise. [21:47:44] Ah [21:48:15] Mw-core could probably be dropped from 5.3 tests infact im shocked it hasnt [21:48:28] https://github.com/wikimedia/integration-config/search?utf8=✓&q=precise&type=Code [21:48:45] Zppix: The point is we still need PHP 5.3 for REL1_23 [21:48:53] Which is the only reason it was kept around [21:48:54] hashar i see one reference to precise for zuul. Probaly want to update that to either jessie or trusty. [21:49:14] Reedy 1.23 support gets dropped in a few months :) [21:49:21] Can't come soon enough [21:49:47] LOL [21:50:00] Why the hell do we support 1.23 still infact i dont even think anyone uses 1.23 (maybe execpt .onion sites) [21:50:09] aude: iirc when I looked at ham crest I basically cloned all the repos then replayed locally what mw-fetch-dev script is doing [21:50:44] LTS [21:50:45] the script is in integration/jenkins.git [21:50:53] Zppix per Reedy [21:50:59] lts is long term support [21:51:04] and invokes tools/composer-dev-args.js to grab the require-dev from mediawiki composer.json [21:51:06] 2 years i think [21:51:13] https://github.com/wikimedia/integration-jenkins/blob/master/bin/mw-fetch-composer-dev.sh ? [21:51:37] yeah [21:51:42] maybe it has an outdated version of mediawiki in some cases? [21:51:57] hashar we could begin removing some php53 tests, for example the composer php 53 tests. which phplint can be covered by the php53lint. [21:51:59] aude: and there is another nodejs script in tools/composer-dev-args.js which process the composer.json file looking for require-dev [21:52:18] I mean wouldnt by now bugs in 1.23 be fixed by now [21:52:27] yeah [21:52:29] aude: that outputs a list of dependencies that are feed to composer to install them. That part apparently works since composer shows it is installing ham crest [21:52:43] Atleast the major ones that would be noticable [21:52:45] aude: maybe the autoloader generation is screwed up [21:52:56] Zppix yes but there are a ton of bugs [21:53:08] aude: I really like the idea Tim0 came up with. it is smart [21:53:12] maybe [21:53:18] some un noticable and some introduced by other changes. [21:53:45] Whats saucelabs? [21:54:41] Zppix i presume browser tests [21:54:43] for ci [21:55:02] It must be new because they werent there yesterday [21:55:09] Zppix where? [21:55:32] Jenkins on integration site [21:57:55] i see none here https://integration.wikimedia.org/ci/ [21:58:09] Click on the default tab [21:58:18] Then scroll all the way down [21:58:23] Its an instance [21:59:02] Zppix oh [21:59:02] https://integration.wikimedia.org/ci/computer/saucelabs-01/ [21:59:29] the label says its Browser Tests, thats zelikof :) [21:59:33] ah yeah [21:59:40] woops wrong spelling zeljkof [21:59:41] Zppix: sorry you asked about them earlier today iirc [21:59:42] so [21:59:59] saucelabs-XXX instances are Jessie instances [21:59:59] with just ruby installed [22:00:05] they are running the jobs selenium-* [22:00:15] those ones are triggered on a daily basic automatically by Jenkins [22:00:29] the intent is to migrate those jobs (and the slaves) to a new Jenkins instance [22:02:08] aude: do fill a task potentially. I guess we can hack on it tomorrow with addshore.. [22:02:16] not a good time for me tonight :/ [22:02:26] hashar this https://phabricator.wikimedia.org/T145669#3038697 is what twentyafterfour found :). We could experiment with that with docker :) [22:02:33] (ci) [22:05:12] it is too early for that really [22:05:26] hashar: https://phabricator.wikimedia.org/T158674 [22:05:29] thanks [22:05:37] i'll try to poke more also and am around tomorrow [22:05:40] paladox: have a look at https://github.com/pearsontechnology/deployment-pipeline-jenkins-plugin [22:06:02] hashar: you should sleep eh? [22:06:16] paladox: apparently lets you spawn several instances, set them up somehow and then run whatever commands against the swarm of containers you have spawned [22:07:06] Zppix: you are right yes :} My wife and I are trying to figure out how to access a local equivalent of Netflix [22:07:22] hashar yep [22:07:43] hashar try amazon prime [22:07:59] paladox: ethically we can not use Amazon. It is banned [22:08:07] can't use netflix either for the same reason [22:08:07] hashar: *cough* tor *cough* [22:08:09] Oh why is it banned? [22:08:16] TAXES EVASION !!!!!!!! [22:08:27] paladox: media copyright prob [22:08:29] also known has fiscal optimization [22:08:30] hashar: What if you use someone elses account, ala you're not paying for it :P [22:08:41] we actually want to pay! [22:08:42] hashar lolololol [22:08:44] Reedy: they block netflix period [22:08:46] Prob [22:09:06] hashar that may not be blocked soon [22:09:24] if the eu implements the everyone can access the website through out the whole of the eu policy [22:10:22] 10Continuous-Integration-Config, 10Wikidata: Fatal error: Call to undefined function Wikibase\Client\Tests\RecentChanges\both() - on jenkins - https://phabricator.wikimedia.org/T158674#3044812 (10hashar) I invoke @Addshore and @Aleksey_WMDE since that is ham crest related. Gotta reproduce and figure out why h... [22:11:11] hashar http://variety.com/2016/biz/global/netflix-amazon-france-1201849106/ [22:11:22] though thats from last year. [22:12:13] 10Continuous-Integration-Config, 10Wikidata: Fatal error: Call to undefined function Wikibase\Client\Tests\RecentChanges\both() - on jenkins - https://phabricator.wikimedia.org/T158674#3044821 (10hashar) [22:12:30] (03PS2) 10Hashar: Capture composer.lock after mw-fetch-composer-dev [integration/config] - 10https://gerrit.wikimedia.org/r/339051 (https://phabricator.wikimedia.org/T158674) [22:12:39] aude: did some basic triage. Will try to reproduce tomorrow [22:13:02] paladox: what I meant by blocking is that we refuse to use Amazon or Netflix [22:13:10] Oh [22:13:46] hashar i doint think there are any other services. [22:13:59] In the UK we also have things like NowTV [22:14:02] Orange has its own one http://go.ocs.fr/ :) [22:14:03] hashar i have sky go, amazon prime and used to have netflix for free [22:14:04] Sky do some on deand stuff [22:14:16] 10Continuous-Integration-Config, 10Wikidata, 13Patch-For-Review: Fatal error: Call to undefined function Wikibase\Client\Tests\RecentChanges\both() - on jenkins - https://phabricator.wikimedia.org/T158674#3044834 (10aude) hamcrest-php and hamcrest-html-matchers are dev dependencies of mediawiki core. Jenkin... [22:14:24] hashar orange is no more here, it's now ee. [22:14:33] and my ISP ( free.fr ) offers dozens of VOD systems [22:14:33] Reedy yeh, all the new content + sky q. [22:14:43] hashar: thanks [22:14:44] guess I will just relocate to the UK [22:14:51] oh, and lol [22:15:04] hashar: just stay the heck away from the USA [22:15:10] hashar you get freeview. [22:15:10] Or wait 8 years [22:15:23] Zppix why 8 years? [22:15:28] aude: I guess I will pick https://gerrit.wikimedia.org/r/#/c/338940/ , run the few CI slave scripts to inject the require-dev and then dig in the autoloader [22:15:50] paladox: new potus [22:15:57] hashar now they have freeview on demand [22:16:05] Zppix oh [22:16:16] and eventually ocs.fr relies on SilverLight !!!!!!!!!!!!!!!! [22:16:25] Zppix we have a new pm from last year. [22:16:29] hashar lol, sky go used to rely on that. [22:16:30] that is like the first time I see a website using Microsoft Silverlight. [22:16:46] hashar i have seen a ton of website like amazon prime, sky go. [22:17:07] hashar: ok [22:18:28] hashar, does france have free on demand movies / tv shows? [22:21:04] note sue [22:21:12] we still have black & white tv you know [22:21:40] oh my f**** [22:22:02] "Since version 45, Chrome is no more compatible with Silverlight" [22:22:06] which is good for Chrome [22:22:12] but terrible for that VOD provider [22:22:14] :} [22:22:29] hashar lololol [22:22:33] tell that to my tv provider [22:22:44] Silverlight people still use that? [22:22:49] Come on people its 2017 [22:23:07] Zppix it's secure. [22:23:28] Zppix it's all down to the content producers. [22:25:37] paladox: html5 is pretty secure [22:27:06] Zppix yes, thats why companys are waiting till the last minute [22:28:32] and I can't unsubscribe from the service bah [22:28:37] 12€ wasted :D [22:28:46] anyway have a good night and see you later! [22:28:49] hashar what service? [22:28:58] hashar: go to sleep your too tired [22:28:59] PROBLEM - Puppet run on deployment-mira is CRITICAL: CRITICAL: 50.00% of data above the critical threshold [0.0] [22:29:00] paladox: the Orange VOD system :) [22:29:18] hashar just do what everyone does with mine, threaten to leave and they give you the best deals [22:29:43] will just kindly ask them to cancel the subscription since SilverLight is crap / does not work [22:30:06] ok [22:33:52] 10Continuous-Integration-Infrastructure, 06Release-Engineering-Team: Depool precise jenkins instances - https://phabricator.wikimedia.org/T158652#3044885 (10Dzahn) [22:33:54] 10Continuous-Integration-Infrastructure, 06Release-Engineering-Team: Depool precise jenkins instances - https://phabricator.wikimedia.org/T158652#3042963 (10Dzahn) [22:33:57] 10Continuous-Integration-Infrastructure, 06Release-Engineering-Team: Depool precise jenkins instances - https://phabricator.wikimedia.org/T158652#3042963 (10Dzahn) [22:35:14] hashar you could always use safari? Or what about chrome 44? [22:46:41] 10Gerrit, 07LDAP: underscores in usernames are not recognized - https://phabricator.wikimedia.org/T50774#508070 (10Paladox) This should be fixed in https://gerrit-review.googlesource.com/#/c/94914/ [22:46:53] 10Gerrit, 07LDAP, 07Upstream: underscores in usernames are not recognized - https://phabricator.wikimedia.org/T50774#3044905 (10Paladox) [22:48:31] 10Gerrit, 07LDAP, 07Upstream: underscores in usernames are not recognized - https://phabricator.wikimedia.org/T50774#508070 (10Paladox) Also see https://bugs.chromium.org/p/gerrit/issues/detail?id=2830 [23:16:48] 10Deployment-Systems, 10Tool-Labs-tools-Other, 13Patch-For-Review, 15User-bd808: Jouncebot not joining #wikimedia-operations - https://phabricator.wikimedia.org/T144189#3044990 (10bd808) 05Open>03Resolved a:03bd808 Fixed a long time ago [23:38:59] RECOVERY - Puppet run on deployment-mira is OK: OK: Less than 1.00% above the threshold [0.0]