[00:02:42] 5Gerrit-Migration, 10Wikimedia-Site-Requests: Migrate mediawiki-config to Differential - https://phabricator.wikimedia.org/T131418#2166792 (10Legoktm) [00:05:07] legoktm: I've gotten that wrong for so long :/ thanks [00:05:17] np [00:05:35] legoktm: are you in israel? [00:06:13] thcipriani: around? I accidentally deleted your patch regarding public key [00:06:27] I'm fixing it but I need to know what was the path again [00:06:45] * thcipriani looks [00:07:51] (I add it my patch in puppet/operations) [00:08:06] btw: It's working, I'm doing minor issues right now [00:08:16] puppet:///private/ssh/tin/servicedeploy_rsa.pub [00:08:25] awesome [00:08:26] thanks [00:08:30] nice! glad to heard everything is working :) [00:08:34] greg-g: So... deploying a new extension on beta, that will only affect users in the Staff group. Should I make a phab ticket outlining all that first? Or should I just deploy and ask forgiveness? [00:09:09] (I clearly don't do this enough to remember what the process currently is) [00:09:53] on the staff page? [00:10:01] oh staff group [00:10:02] greg-g: San Jose :P [00:10:18] Yeah Staff (global-) group [00:10:25] csteipp: did you get a security review? :P [00:10:38] Dang it! Knew I was forgetting something... [00:11:03] so after you do that :) jfdi pretty much. [00:11:20] cool :) [00:11:21] obviously, I like knowing so asking/filing a task with the plan is good [00:11:30] Last lame question, how do I log to the releng SAL? [00:11:32] #beta-cluster-infrastructure [00:11:40] do a !log in here [00:14:54] 10Beta-Cluster-Infrastructure, 6Security-Team: Install Ex:OATH to beta - https://phabricator.wikimedia.org/T131420#2166800 (10csteipp) [00:16:14] 10Beta-Cluster-Infrastructure, 6Security-Team: Install Ex:OATH to beta - https://phabricator.wikimedia.org/T131420#2166800 (10greg) +1 [00:16:27] 5Gerrit-Migration, 10Differential: Find way to use Differential with plain git (i.e.: without requiring arc) - https://phabricator.wikimedia.org/T127#2166815 (10Paladox) @mmodell hi question: would it be possible to have arc installed on the server and when someone uses there normal git install and push would... [00:16:30] !log created oathauth_users table on centralauth db in beta [00:16:33] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [00:24:59] 5Gerrit-Migration, 10Differential: Find way to use Differential with plain git (i.e.: without requiring arc) - https://phabricator.wikimedia.org/T127#2166822 (10Paladox) Alsoit seems that repos are going to start being migrated to diffusion. See https://phabricator.wikimedia.org/T130418 All the migration is m... [00:37:55] 5Gerrit-Migration, 10Wikimedia-Site-Requests: Migrate mediawiki-config to Differential - https://phabricator.wikimedia.org/T131418#2166731 (10Luke081515) I can take the replication part (replication to github) [00:40:27] 5Gerrit-Migration, 10MediaWiki-Vagrant: Migrate mediawiki-vagrant to Differential - https://phabricator.wikimedia.org/T131419#2166756 (10Luke081515) Same like here: T131418#2166828 [00:41:59] Yippee, build fixed! [00:42:00] Project selenium-Flow » firefox,beta,Linux,,contintLabsSlave && UbuntuTrusty build #10: 09FIXED in 25 min: https://integration.wikimedia.org/ci/job/selenium-Flow/BROWSER=firefox,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=Linux,VERSION=,label=contintLabsSlave%20&&%20UbuntuTrusty/10/ [00:44:57] csteipp: oathauth is the 2FA thing? [00:45:10] Luke081515: Yep [00:45:29] csteipp: Already testable, or is this just the DB? [00:46:16] Luke081515: should be available soon. [00:46:26] ok :D [00:46:37] Luke081515: I'm only turning it on for Staff right now, but if there's another group who would like to try it out, happy to add the right. [00:47:07] ok [01:42:02] RECOVERY - Puppet run on deployment-ores-web is OK: OK: Less than 1.00% above the threshold [0.0] [01:48:06] PROBLEM - Puppet run on deployment-ores-web is CRITICAL: CRITICAL: 50.00% of data above the critical threshold [0.0] [03:22:26] PROBLEM - Puppet run on deployment-sca01 is CRITICAL: CRITICAL: 60.00% of data above the critical threshold [0.0] [04:00:28] !log manually rebooted deployment-sca01 [04:00:32] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [04:02:02] RECOVERY - Puppet run on deployment-ores-redis is OK: OK: Less than 1.00% above the threshold [0.0] [04:24:10] Yippee, build fixed! [04:24:11] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-windows_7-internet_explorer-9-sauce build #768: 09FIXED in 32 min: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-windows_7-internet_explorer-9-sauce/768/ [04:44:34] 5Gerrit-Migration, 10Differential: Find way to use Differential with plain git (i.e.: without requiring arc) - https://phabricator.wikimedia.org/T127#2166961 (10greg) >>! In T127#2166822, @Paladox wrote: > Alsoit seems that repos are going to start being migrated to diffusion. See https://phabricator.wikimedia... [04:55:05] 10Beta-Cluster-Infrastructure, 6Revision-Scoring-As-A-Service, 10ores: Setup ORES service in beta cluster - https://phabricator.wikimedia.org/T130404#2166962 (10Ladsgroup) [05:27:08] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-windows_7-internet_explorer-11-sauce build #744: 04FAILURE in 25 min: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-windows_7-internet_explorer-11-sauce/744/ [05:38:24] 10Beta-Cluster-Infrastructure, 6Revision-Scoring-As-A-Service, 10ores: Setup ORES service in beta cluster - https://phabricator.wikimedia.org/T130404#2167013 (10Ladsgroup) Now, it's live in https://ores-beta.wmflabs.org/ - The web node is: deployment-ores-web.deployment-prep.eqiad.wmflabs - The worker... [05:54:04] https://phabricator.wikimedia.org/T130404#2167013 [05:54:08] thcipriani|afk: ^ [05:54:18] ostriches: ^ [06:33:52] Yippee, build fixed! [06:33:53] Project browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce build #66: 09FIXED in 6 min 52 sec: https://integration.wikimedia.org/ci/job/browsertests-VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox-sauce/66/ [06:37:57] Project beta-scap-eqiad build #96235: 04FAILURE in 13 min: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/96235/ [06:42:06] Project beta-scap-eqiad build #96236: 04STILL FAILING in 2 min 24 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/96236/ [06:47:06] Project beta-scap-eqiad build #96237: 04STILL FAILING in 2 min 21 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/96237/ [06:50:03] Yippee, build fixed! [06:50:03] Project browsertests-Flow-en.wikipedia.beta.wmflabs.org-linux-firefox-monobook-sauce build #775: 09FIXED in 25 min: https://integration.wikimedia.org/ci/job/browsertests-Flow-en.wikipedia.beta.wmflabs.org-linux-firefox-monobook-sauce/775/ [06:58:49] Yippee, build fixed! [06:58:49] Project beta-scap-eqiad build #96238: 09FIXED in 4 min 2 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/96238/ [07:18:31] hi, I want to activate ferm firewall rules on terbium on Monday morning Euro. that should not have any impact on running scripts. but if there's something terribly critical where we can't take any risks of interrupting it, please get in touch with me [07:19:11] it's already enabled on wasat (the codfw equivalent of terbium) w/o any problems [07:32:08] Yippee, build fixed! [07:32:09] Project browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-windows_8-internet_explorer-10-sauce build #367: 09FIXED in 23 min: https://integration.wikimedia.org/ci/job/browsertests-MultimediaViewer-en.wikipedia.beta.wmflabs.org-windows_8-internet_explorer-10-sauce/367/ [07:32:31] Yippee, build fixed! [07:32:31] Project browsertests-MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-firefox-sauce build #1032: 09FIXED in 31 min: https://integration.wikimedia.org/ci/job/browsertests-MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-firefox-sauce/1032/ [07:35:36] 10Browser-Tests-Infrastructure, 6Release-Engineering-Team, 7Epic, 13Patch-For-Review, and 2 others: Fix scenarios that fail at en.wikipedia.beta.wmflabs.org or do not run them daily - https://phabricator.wikimedia.org/T94150#2167185 (10Jdlrobson) [07:35:38] 10Browser-Tests-Infrastructure, 10Reading-Web, 13Patch-For-Review, 5WMF-deploy-2016-03-22_(1.27.0-wmf.18), 5WMF-deploy-2016-04-05_(1.27.0-wmf.20): Fix MobileFrontend scenarios that fail at en.wikipedia.beta.wmflabs.org or do not run them daily - https://phabricator.wikimedia.org/T94156#2167182 (10Jdlrobso... [08:19:04] PROBLEM - Free space - all mounts on deployment-fluorine is CRITICAL: CRITICAL: deployment-prep.deployment-fluorine.diskspace._srv.byte_percentfree (<40.00%) [09:19:24] 10Continuous-Integration-Infrastructure, 5Gerrit-Migration, 3releng-201516-q4, 6Operations, and 4 others: Make sure phab can talk to gearman and nodepool instances can talk to phabricator - https://phabricator.wikimedia.org/T131375#2167638 (10hashar) >>! In T131375#2165682, @mmodell wrote: > @hashar: I don... [09:21:23] (03CR) 10Hashar: [C: 032] [IndexFunction] Add npm test and composer-test [integration/config] - 10https://gerrit.wikimedia.org/r/280775 (owner: 10Paladox) [09:22:34] (03Merged) 10jenkins-bot: [IndexFunction] Add npm test and composer-test [integration/config] - 10https://gerrit.wikimedia.org/r/280775 (owner: 10Paladox) [09:24:58] RECOVERY - Host integration-trusty-1026 is UP: PING OK - Packet loss = 0%, RTA = 0.84 ms [09:25:15] 5Gerrit-Migration, 10Wikimedia-Site-Requests: Migrate mediawiki-config to Differential - https://phabricator.wikimedia.org/T131418#2167665 (10greg) [09:30:59] 10Deployment-Systems, 10Gerrit, 10ReleaseTaggerBot, 7WorkType-NewFunctionality: Deployment status indicator for gerrit patches - https://phabricator.wikimedia.org/T88136#2167668 (10greg) a:5hashar>3None [09:34:54] PROBLEM - Host integration-trusty-1026 is DOWN: CRITICAL - Host Unreachable (10.68.17.98) [09:42:24] Yippee, build fixed! [09:42:24] Project browsertests-Flow-en.wikipedia.beta.wmflabs.org-linux-chrome-monobook-sauce build #766: 09FIXED in 21 min: https://integration.wikimedia.org/ci/job/browsertests-Flow-en.wikipedia.beta.wmflabs.org-linux-chrome-monobook-sauce/766/ [09:54:38] (03PS1) 10Addshore: Add tinajohnson to jenkins +2 test pipe [integration/config] - 10https://gerrit.wikimedia.org/r/280845 [09:54:46] hashar: alive? :D [09:54:51] also, not at the hackathon? [09:55:44] (03CR) 1001tonythomas: [C: 031] Add tinajohnson to jenkins +2 test pipe [integration/config] - 10https://gerrit.wikimedia.org/r/280845 (owner: 10Addshore) [09:58:00] addshore: indeed I am not attending the hackathon [09:58:09] sad times :( [09:58:17] how do you feel about merging the patch above for me? :D [10:04:03] (03CR) 10Hashar: [C: 032] Add tinajohnson to jenkins +2 test pipe [integration/config] - 10https://gerrit.wikimedia.org/r/280845 (owner: 10Addshore) [10:04:05] addshore: doing :) [10:04:55] (03Merged) 10jenkins-bot: Add tinajohnson to jenkins +2 test pipe [integration/config] - 10https://gerrit.wikimedia.org/r/280845 (owner: 10Addshore) [10:05:04] :) hashar thanks! :D [10:05:55] addshore: and deployed :-} [10:06:56] 10scap, 10Analytics-EventLogging, 6Analytics-Kanban, 13Patch-For-Review, 10Scap3 (Scap3-Adoption-Phase1): Use scap3 to deploy eventlogging/eventlogging - https://phabricator.wikimedia.org/T118772#2167804 (10mobrovac) [10:11:53] 10Continuous-Integration-Infrastructure: database disk image is malformed in test mwext-qunit-composer - https://phabricator.wikimedia.org/T124611#2167866 (10hashar) 5Open>3Resolved a:3hashar No clue what it was, but apparently does not occur anymore [10:19:45] 10Continuous-Integration-Config, 13Patch-For-Review: Have extensions with dependencies use the generic mwext-testextension-* job - https://phabricator.wikimedia.org/T96690#2167899 (10hashar) Well handled @Paladox . Thank you very much. [10:22:09] 10Continuous-Integration-Infrastructure: Consolidate jobs to test entry points - https://phabricator.wikimedia.org/T91996#2167911 (10hashar) [10:22:11] 10Continuous-Integration-Config: Replace project-specific "{name}-thing" jobs with generic "thing" ones - https://phabricator.wikimedia.org/T91997#2167908 (10hashar) 5Open>3Resolved a:3hashar I am claiming this task to be solved after a looooong period of jobs refactoring. There are still some jobs varyin... [10:26:46] (03PS1) 10Hashar: [Offline] archive extension [integration/config] - 10https://gerrit.wikimedia.org/r/280856 [10:28:36] (03CR) 10Hashar: [C: 032] [Offline] archive extension [integration/config] - 10https://gerrit.wikimedia.org/r/280856 (owner: 10Hashar) [10:29:23] (03Merged) 10jenkins-bot: [Offline] archive extension [integration/config] - 10https://gerrit.wikimedia.org/r/280856 (owner: 10Hashar) [10:32:01] (03PS1) 10Hashar: Name the archived gate queue [integration/config] - 10https://gerrit.wikimedia.org/r/280857 [10:34:27] (03CR) 10Hashar: [C: 032] Name the archived gate queue [integration/config] - 10https://gerrit.wikimedia.org/r/280857 (owner: 10Hashar) [10:35:10] (03Merged) 10jenkins-bot: Name the archived gate queue [integration/config] - 10https://gerrit.wikimedia.org/r/280857 (owner: 10Hashar) [11:11:25] 10Beta-Cluster-Infrastructure, 6Commons, 10wikimaps-warper: Thumbnails and images seem to be 503 erroring on Beta Commons - https://phabricator.wikimedia.org/T131472#2168004 (10Chippyy) [11:23:25] !log Freed 4.5G on deployment-fluorine:/srv/mw-log by deleting wfDebug.log [11:23:30] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [11:30:20] 10Beta-Cluster-Infrastructure, 6Commons, 10MediaWiki-File-management, 6Multimedia: Thumbnail generation should happen via the same setup in the beta cluster and in production (tracking) - https://phabricator.wikimedia.org/T84950#2168051 (10Tgr) Video also seems broken ([[http://commons.wikimedia.beta.wmfla... [11:40:42] 10Continuous-Integration-Infrastructure, 7Technical-Debt, 7Tracking: All repositories should pass jshint test (tracking) - https://phabricator.wikimedia.org/T62619#2168116 (10Paladox) [11:41:45] 10Continuous-Integration-Infrastructure, 7Technical-Debt, 7Tracking: All repositories should pass jshint test (tracking) - https://phabricator.wikimedia.org/T62619#658851 (10Paladox) [11:42:28] 10Continuous-Integration-Infrastructure, 7Technical-Debt, 7Tracking: All repositories should pass jshint test (tracking) - https://phabricator.wikimedia.org/T62619#658851 (10Paladox) [11:42:59] 10Continuous-Integration-Infrastructure, 7Technical-Debt, 7Tracking: All repositories should pass jshint test (tracking) - https://phabricator.wikimedia.org/T62619#2168137 (10Paladox) [11:52:13] Hi. A small note about CI: the mediawiki-extensions-qunit job fail for core. https://integration.wikimedia.org/ci/job/mediawiki-extensions-qunit/37329/console 11:41:45 npm ERR! tar.unpack untar error /mnt/home/jenkins-deploy/.npm/socket.io-client/1.4.5/package.tgz [11:54:41] No, other tests on core succeed. [11:54:46] Dereckson: Known problem upgrading to nodepool fixes it. [11:55:54] Dereckson: If it continues to fail on your next recheck we would need to look into it since all other cases after a recheck fixed it. [11:56:08] Yours looks like a custom package problem from one of the packages. [11:56:17] let's try a second recheck (it failed twice) [11:56:38] arf the edit conflict on Gerrit [11:59:03] Dereckson: Seems to be working now, see https://integration.wikimedia.org/ci/job/mediawiki-extensions-qunit/37332/console please. [12:01:05] Thanks. [12:08:19] PROBLEM - Puppet run on deployment-imagescaler01 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [0.0] [12:35:47] 10Continuous-Integration-Infrastructure: integration-slave-trusty-1025 has old puppet-lint version - https://phabricator.wikimedia.org/T131485#2168437 (10JanZerebecki) [12:38:54] 10Continuous-Integration-Infrastructure: integration-slave-trusty-1025 has old puppet-lint version - https://phabricator.wikimedia.org/T131485#2168471 (10hashar) 5Open>3Resolved a:3hashar Same on integration-slave-trusty-1024.integration.eqiad.wmflabs . Did an apt-get upgrade on both slaves and they now h... [12:41:34] 7Blocked-on-RelEng, 10Continuous-Integration-Infrastructure, 6Labs, 10Tool-Labs, 13Patch-For-Review: debian-glue tries to fetch obsolete package - https://phabricator.wikimedia.org/T125999#2168493 (10hashar) 5Open>3Resolved a:3hashar Yup looks fine now. [12:45:37] 10Continuous-Integration-Infrastructure: integration-slave-trusty-1025 has old puppet-lint version - https://phabricator.wikimedia.org/T131485#2168504 (10JanZerebecki) [12:45:57] (03PS4) 10Hashar: [parsoidsvc] Update npm tests [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (owner: 10Paladox) [12:46:08] (03PS5) 10Hashar: [parsoidsvc] Update npm tests [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (https://phabricator.wikimedia.org/T126992) (owner: 10Paladox) [12:50:08] (03CR) 10Hashar: [parsoidsvc] Update npm tests (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (https://phabricator.wikimedia.org/T126992) (owner: 10Paladox) [12:50:13] (03PS6) 10Hashar: [parsoidsvc] Update npm tests [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (https://phabricator.wikimedia.org/T126992) (owner: 10Paladox) [12:52:03] (03CR) 10Hashar: [C: 032] "Excellent Paladox, I have just removed the 'throttle one per node' parameter which is no more needed on Nodepool instances (they have a si" [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (https://phabricator.wikimedia.org/T126992) (owner: 10Paladox) [12:53:01] (03CR) 10Paladox: "@Hashar thanks, and thanks for merging my patches today. I will ask on irc to see if they want npm 0.10 to be dropped to." [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (https://phabricator.wikimedia.org/T126992) (owner: 10Paladox) [12:53:40] subbu: Hi, question would you like us to drop npm 0.10 from parsoid too. Per hashar question at https://gerrit.wikimedia.org/r/#/c/279529/ [12:53:57] hashar: HI [12:54:03] O/ [12:54:18] hashar: Thanks for merging my patches. I asked the question about 0.10 above. [12:54:33] (03Merged) 10jenkins-bot: [parsoidsvc] Update npm tests [integration/config] - 10https://gerrit.wikimedia.org/r/279529 (https://phabricator.wikimedia.org/T126992) (owner: 10Paladox) [12:55:43] hashar: Could you review https://gerrit.wikimedia.org/r/#/c/279529/ please, its for visualeditor switching to npm 4.3. [12:56:24] hashar: Also could this be merged https://gerrit.wikimedia.org/r/#/c/280664/ so when i finished converting mediawiki extensions we can see about doing it for all repos. [12:56:26] please [12:56:55] hashar: Also this https://gerrit.wikimedia.org/r/#/c/262432/ will fix the visualeditor dirty submodule. [12:57:06] and reusing code. [12:58:22] waiting for parsoid tests with Node 4.3 to complete -} [12:59:08] PROBLEM - Free space - all mounts on deployment-fluorine is CRITICAL: CRITICAL: deployment-prep.deployment-fluorine.diskspace._srv.byte_percentfree (<44.44%) [13:00:41] hashar: Ok, thanks. [13:01:31] hashar: Also you know oojs when we had the problem with phantomjs it seems they have now switched there download links to there location on github. :) also seems they support caching now. [13:02:47] hashar: It says jsduck not avalible. [13:03:08] hashar: Would we be able to include jsduck for now and then start migrating repos to use rake-jessie [13:08:05] 5Continuous-Integration-Scaling, 10Parsoid, 6Services, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate Parsoid CI jobs from node 0.8/0.10 to 4.3 - https://phabricator.wikimedia.org/T126992#2168624 (10hashar) https://gerrit.wikimedia.org/r/#/c/279529/ Drops the node 0.8 jobs and add the 4.3 jobs.... [13:09:26] hashar: How would we switch parsoid to https://phabricator.wikimedia.org/T109005 since i carn't find any example or any other repo doing that. [13:13:49] paladox: i am figuring out a hack [13:14:29] hashar: Oh, thanks. Im having a look too, by hack do you mean how to install on the server or hack as in how to run it on rake. [13:17:56] paladox: will get 'npm doc' to install jsduck from ruby gems and run it under the environment setup [13:18:00] something like [13:18:02] bundle install [13:18:05] bundle exec jsduck [13:18:19] hashar: oh ok. [13:19:02] parsoid has so many deps [13:19:27] hashar: Yep [13:22:07] hashar: We can apply that to MobileFrontend if it works with parsoid. [13:25:26] yeah [13:33:01] hashar: It seems to say Bundler::GemspecError: Could not read gem at /var/lib/gems/2.1.0/cache/dimensions-1.2.0.gem. It may be corrupted. [13:33:34] oh stupid Debian [13:33:35] seriously [13:33:51] tries to write to /var/lib/gems/2.1.0/cache/ [13:33:55] which is ... root owned [13:33:56] :( [13:34:08] hashar: Oh, we may want todo then [13:34:08] bundle install --clean --path ../vendor/bundle [13:34:35] it is Debian being ridiculously annoying [13:34:35] then bundle exec ../vendor/bundle/jsduck [13:34:40] hashar: Yep. [13:35:40] hashar: Also happend to npm 0.10 [13:35:44] https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-0.10/4366/console [13:36:43] 10Continuous-Integration-Infrastructure, 6Operations, 10Traffic, 13Patch-For-Review, 7Varnish: Make CI run Varnish VCL tests - https://phabricator.wikimedia.org/T128188#2168733 (10ema) [13:43:43] paladox. no please. we need node 0.10. [13:44:11] subbu: Ok it was only a question we were never going to remove it without asking you first and waiting for your reply. [13:44:28] understood. thanks for checking. [13:44:44] subbu: We are trying to fix it since we need to use gem and bundle to install jsduck. So currently the tests will fail. [13:45:13] hashar: Seems some packages need npm 2.x [13:45:21] 13:41:49 npm WARN engine joi@6.10.0: wanted: {"node":">=0.10.40","npm":">=2.0.0"} (current: {"node":"4.3.0","npm":"1.4.21"}) [13:45:31] paladox: feel free to fill bug under the task that request npm 2.x to be installed [13:45:39] but I think npm magically upgrade itself [13:45:51] hashar: What about installing nvm. [13:46:50] hashar: Described at https://phabricator.wikimedia.org/T124474#2143837 [13:47:30] hashar: We could add the updating npm script to the npm test. It may add a few secs to the test but its worth having an updated npm that is still supported. [13:52:48] hashar: Maybe we just want to do npm install npm@2 just before doing npm install. Since that would be the temp way. [13:57:27] hashar: It says this [13:57:28] Bundle defaults to install to GEM_HOME which on Debian is the system [14:01:53] 10Continuous-Integration-Infrastructure: integration-slave-trusty-1025 has old puppet-lint version - https://phabricator.wikimedia.org/T131485#2168892 (10Dzahn) thank you! i had noticed the "invalid option: --no-puppet_url_without_modules-check " on some checks but then if you did a 'recheck" it did not happen.... [14:05:40] !log integration: removed sudo policy permitting sudo as root for any member of the project, which included jenkins-deploy user [14:05:45] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [14:06:15] !log integration: removed sudo policy permitting sudo as any member of the project for any member of the project, which included jenkins-deploy user [14:06:20] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [14:08:49] (03PS1) 10Paladox: Update npm to 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280912 [14:08:55] hashar: ^^ [14:10:49] (03CR) 10Paladox: "It only adds a few secs on but it is worth it having a supported and secured package instead of a non supported and security risked versio" [integration/config] - 10https://gerrit.wikimedia.org/r/280912 (owner: 10Paladox) [14:12:40] paladox: hold on [14:12:59] paladox: if it takes me an hour to process an idea and you get one every 10 minutes I can't follow ;-} [14:13:06] still playing with parsoid/npm/jsduck [14:13:14] and I think I got a solution!!! [14:16:48] hashar oh sorry and ok [14:17:19] trick would be to inject the env variable BUNDLE_PATH='~/.bundle' [14:17:28] so bundle install will install the gems there [14:18:48] yep [14:21:00] lets hack something [14:21:12] 10Beta-Cluster-Infrastructure, 6Commons, 10wikimaps-warper: Thumbnails and images seem to be 503 erroring on Beta Commons - https://phabricator.wikimedia.org/T131472#2169028 (10Krenair) [14:21:14] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2169029 (10Krenair) [14:21:59] hashar: Yep [14:22:17] hashar: :) [14:31:35] hashar: We need to update parsoid npm 4.3 name to parsoidsvc-{repository}-npm-node-4.3 so [14:31:35] elif job.name.endswith('npm-node-4.3'): [14:31:35] offline_when_complete(item, job, params) [14:31:37] can work [14:33:34] (03PS1) 10Hashar: Pass BUNDLE_PATH to Nodepool jobs [integration/config] - 10https://gerrit.wikimedia.org/r/280919 (https://phabricator.wikimedia.org/T109005) [14:33:40] (03PS1) 10Paladox: Update parsoidsvc test to have npm 4.3 named like npm-node-4.3 not npm-4.3 [integration/config] - 10https://gerrit.wikimedia.org/r/280920 [14:33:44] hashar: ^^ [14:34:58] yeahhhh [14:35:00] good one :-) [14:35:19] should really rename all those jobs to use a '-nodepool' suffix [14:35:36] (03CR) 10Hashar: [C: 032] Update parsoidsvc test to have npm 4.3 named like npm-node-4.3 not npm-4.3 [integration/config] - 10https://gerrit.wikimedia.org/r/280920 (owner: 10Paladox) [14:36:16] hashar: Yeh, do you want me to do that once we have done our testing and fixed jobs needing jsduck to use bundle and updating npm. [14:36:22] hashar: Also thanks [14:37:38] (03CR) 10jenkins-bot: [V: 04-1] Update parsoidsvc test to have npm 4.3 named like npm-node-4.3 not npm-4.3 [integration/config] - 10https://gerrit.wikimedia.org/r/280920 (owner: 10Paladox) [14:37:44] (03PS2) 10Hashar: Pass BUNDLE_PATH to Nodepool jobs [integration/config] - 10https://gerrit.wikimedia.org/r/280919 (https://phabricator.wikimedia.org/T109005) [14:38:45] (03CR) 10Hashar: [C: 032] "I have created the jobs with JJB. Lets try again" [integration/config] - 10https://gerrit.wikimedia.org/r/280920 (owner: 10Paladox) [14:39:07] hashar: Thanks [14:40:48] (03Merged) 10jenkins-bot: Update parsoidsvc test to have npm 4.3 named like npm-node-4.3 not npm-4.3 [integration/config] - 10https://gerrit.wikimedia.org/r/280920 (owner: 10Paladox) [14:41:54] (03CR) 10Hashar: [C: 032] Pass BUNDLE_PATH to Nodepool jobs [integration/config] - 10https://gerrit.wikimedia.org/r/280919 (https://phabricator.wikimedia.org/T109005) (owner: 10Hashar) [14:44:10] (03Merged) 10jenkins-bot: Pass BUNDLE_PATH to Nodepool jobs [integration/config] - 10https://gerrit.wikimedia.org/r/280919 (https://phabricator.wikimedia.org/T109005) (owner: 10Hashar) [14:50:05] hashar: It is saying this [14:50:06] 14:47:11 npm ERR! cb() never called! [14:50:06] 14:47:11 npm ERR! not ok code 0 [14:50:10] https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/1/console [14:51:34] paladox: yeah I have no clue what it is [14:52:03] hashar: Yeh me too. I doint know why it is doing it now. It was working when you first uploaded the patch. [14:55:31] hashar: http://stackoverflow.com/questions/15393821/npm-err-cb-never-called [14:55:39] hashar: Try clearing the cache. [14:55:59] Maybe related to because it wasent running [14:55:59] elif job.name.endswith('npm-node-4.3'): [14:55:59] offline_when_complete(item, job, params) [14:56:00] there is no cache [14:56:02] at the end [14:56:04] hashar: Oh [14:58:56] hacked the job to run npm install --verbose [14:59:08] https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/4/console [15:00:04] hashar: Ah thanks. We can now find were it comes from. [15:00:05] :0 [15:00:06] :) [15:00:21] hopefully! [15:03:50] hashar: Updating npm to 2.x worked for me [15:03:52] https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/5/console [15:03:54] See ^^ [15:04:08] hashar: https://gerrit.wikimedia.org/r/#/c/280907/ [15:05:23] paladox: what is da difference? [15:05:39] OH [15:05:40] npm 2 [15:05:41] :( [15:05:44] hashar: Npm on nodepool is running npm 1.x so may have more bugs then npm 2. [15:05:50] Oh you noticed [15:07:26] 10Continuous-Integration-Infrastructure, 6Operations, 6Services: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#2169270 (10hashar) [15:07:29] 10Continuous-Integration-Config, 6Operations, 13Patch-For-Review: Switch CI from jsduck deb package to a gemfile/bundler system - https://phabricator.wikimedia.org/T109005#2169269 (10hashar) [15:07:54] hashar: May be related to 14:46:10 npm WARN engine joi@6.10.0: wanted: {"node":">=0.10.40","npm":">=2.0.0"} (current: {"node":"4.3.0","npm":"1.4.21"}) needing npm 2.x [15:08:05] 10Continuous-Integration-Config, 6Operations, 13Patch-For-Review: Switch CI from jsduck deb package to a gemfile/bundler system - https://phabricator.wikimedia.org/T109005#1537429 (10hashar) And the job fails with `cb() never called` which is due to npm1 :( Hence blocked by T124474. [15:08:52] 10Continuous-Integration-Infrastructure, 6Operations, 6Services: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#2169276 (10Paladox) We will need to temporarily do npm install npm@2 for now until we either package npm 2 or using nvm. [15:09:08] pfff [15:09:14] guess I will get npm2 installed in the image [15:09:29] hashar: Oh, thanks. [15:09:34] hashar: Should we go with npm 3 [15:09:56] If we have to install it in the image, and it is hard to do that should we go with npm 3. [15:10:22] Timo stated Node4 comes with npm 2.x [15:10:56] hashar: Oh, but npm 3 works with nodejs 4.x. [15:11:05] just is not included in it [15:11:22] Since nodejs would need a bump in there version thats why it was bumped to 5.x too. [15:11:32] hashar: Nodejs 4.4 is out [15:11:43] ah [15:11:51] jsduck pass https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/5/console :D [15:11:57] Yeh :) [15:12:16] We go with your patch but it was just a test. I was using yours so i could test. [15:12:21] But it passed :) [15:13:05] thcipriani: HIiiiiiiii good morning! [15:13:05] hashar: Should we also update nodejs to 4.4 as we are updating npm to 2 or 3. [15:13:13] ottomata: hiya [15:13:19] (03Abandoned) 10Paladox: Update npm to 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280912 (owner: 10Paladox) [15:13:22] too early to ask you some questions? :D [15:13:35] 10Continuous-Integration-Config, 6Operations, 13Patch-For-Review: Switch CI from jsduck deb package to a gemfile/bundler system - https://phabricator.wikimedia.org/T109005#2169281 (10hashar) So repositories should get a Gemfile with jsduck and trigger the job `rake-jessie` Alternatively, the npm entry point... [15:13:48] 10Continuous-Integration-Infrastructure, 6Operations, 6Services: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#2169283 (10hashar) [15:13:50] ottomata: heh, never too early to ask. May be too early for me to give cogent answers :) [15:13:50] 10Continuous-Integration-Config, 6Operations, 13Patch-For-Review: Switch CI from jsduck deb package to a gemfile/bundler system - https://phabricator.wikimedia.org/T109005#2169282 (10hashar) [15:14:21] 10Continuous-Integration-Infrastructure, 6Operations, 6Services: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#1957157 (10hashar) [15:14:24] 5Continuous-Integration-Scaling, 10Parsoid, 6Services, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate Parsoid CI jobs from node 0.8/0.10 to 4.3 - https://phabricator.wikimedia.org/T126992#2169284 (10hashar) [15:14:58] 5Continuous-Integration-Scaling, 10Parsoid, 6Services, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate Parsoid CI jobs from node 0.8/0.10 to 4.3 - https://phabricator.wikimedia.org/T126992#2029093 (10hashar) The npm install step fails due to npm 1.x with 'cb() has not been called'. Would need np... [15:15:29] hashar: Probly want to also pass params['BUNDLE_PATH'] = '~/workspace/vendor/bundle' to npm 0.10 test too. [15:15:51] yeah [15:16:51] well actually no [15:17:51] a different value would be needed [15:18:04] ok i ask you answer at will [15:18:14] i'm confused about the $package_name param on scap::target [15:18:24] which gets passed down to the scap3 package provider [15:18:43] https://github.com/wikimedia/operations-puppet/blob/production/modules/scap/lib/puppet/provider/package/scap3.rb#L129 [15:19:25] (03PS1) 10Paladox: Export BUNDLE_PATH to npm and npm-0.10 tests [integration/config] - 10https://gerrit.wikimedia.org/r/280929 [15:19:39] hashar: Oh, what value do we use [15:20:15] (03CR) 10jenkins-bot: [V: 04-1] Export BUNDLE_PATH to npm and npm-0.10 tests [integration/config] - 10https://gerrit.wikimedia.org/r/280929 (owner: 10Paladox) [15:20:44] I am going to drop node4.3 for parsoid [15:20:46] (03PS2) 10Paladox: Export BUNDLE_PATH to npm and npm-0.10 tests [integration/config] - 10https://gerrit.wikimedia.org/r/280929 [15:20:53] blocking it until npm 2.x is provisioned [15:21:37] ottomata: yep, so since the ssh_user needs to be able to create a directory in $(dirname $package_name), if the $package name doesn't contain a '/' we make repo_path "$package_name/$package_name" so that ssh_user can do the ln -sfT "$package_name/${package_name}-cache/revs/[sha1]" "$package_name/$package_name" [15:21:42] what's up with this [repo, repo] join? [15:21:50] hashar: Ok, should we just add npm install npm@2 to npm node 4.3 script [15:21:56] ah [15:21:57] hm [15:22:25] hashar: Do you know when npm 2.x will be provisioned [15:22:28] ok, but what if the repository in gerrit is just 'eventlogging' [15:22:37] but the place you want it to be deployed at is /srv/deployment/eventlogging/eventbus [15:22:42] because it is an instance of the source code? [15:23:13] if package_name is eventlogging/eventbus [15:23:19] how does it find the actual git repository? [15:23:37] (03PS1) 10Hashar: Parsoid: stop triggering Node4.3 job [integration/config] - 10https://gerrit.wikimedia.org/r/280932 (https://phabricator.wikimedia.org/T126992) [15:24:23] OH because it is cloned on tin that way [15:24:26] paladox: or we can get jsduck to run via rake-jessie :D [15:24:27] ottomata: so what it's named in gerrit isn't too important for the targets, it just matters what it's named on...yeah [15:24:31] ayyyy [15:24:39] 6Release-Engineering-Team: MW-1.27.0-wmf.18 deployment blockers - https://phabricator.wikimedia.org/T129691#2169330 (10greg) 5Open>3Resolved [15:24:39] hashar: Yeh we could do that. [15:24:49] hashar: How do we get it to do it. [15:24:51] paladox: meanwhile I am disabling parsoid / node4.3 till npm2.x is availab [15:25:11] (03CR) 10Hashar: [C: 032] Parsoid: stop triggering Node4.3 job [integration/config] - 10https://gerrit.wikimedia.org/r/280932 (https://phabricator.wikimedia.org/T126992) (owner: 10Hashar) [15:25:19] hashar: Ok, do you know when npm 2 will be avalible. Or we can do another hack and do npm install npm@2. [15:25:30] no hack [15:25:35] that should be shipped with a deb package [15:26:02] ok thcipriani thanks, that makes more sense then [15:26:04] (03Merged) 10jenkins-bot: Parsoid: stop triggering Node4.3 job [integration/config] - 10https://gerrit.wikimedia.org/r/280932 (https://phabricator.wikimedia.org/T126992) (owner: 10Hashar) [15:26:30] 5Continuous-Integration-Scaling, 10Parsoid, 6Services, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate Parsoid CI jobs from node 0.8/0.10 to 4.3 - https://phabricator.wikimedia.org/T126992#2169335 (10hashar) 5Open>3stalled stalled until npm 2.x is provisioned T124474 [15:26:34] i'm balancing this scap::source thing from scap::target, but package_name doesn't make so much sense in scap::source [15:26:40] since a 'package_name' comes from the deploy server [15:26:50] although, maybe i shoudl use the same concept [15:26:54] and do $package_name = $title [15:27:01] and base everything else in the define off of package_name [15:27:06] $repository is still configurable [15:27:27] hashar: https://packages.debian.org/sid/nodejs [15:27:55] paladox: I am not going to deal with the pain and suffering of backporting/building a deb package [15:28:03] it is just ridiculously long / tedious and crazy [15:28:09] ottomata: I think that makes sense. I was just thinking about that wrt the mkdir_scap_source_path_${title} in scap::source [15:28:37] hashar: Oh ok. Could we do npm install npm@2 it is easisar and allows us to use an npm release that is supported and secured. [15:28:57] thcipriani: yeah, wasn't sure how else to do that [15:30:22] thcipriani: fyi, i'm going to squash my eventlogging::deployment::target/source commits into one, so i can work on them and we can review them together. right now ::target is separate [15:30:42] hashar [15:30:43] https://www.npmjs.com/package/npm2debian [15:30:53] Or this may make it easisar to convert ^^ [15:31:04] npm is a package so we could do it like that [15:31:19] paladox: ops would never accept that :) [15:31:27] hashar: Oh ok [15:36:16] (03Restored) 10Paladox: Update npm to 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280912 (owner: 10Paladox) [15:36:30] hashar: Ok, i have this patch https://gerrit.wikimedia.org/r/#/c/280912/ [15:37:00] 10Browser-Tests-Infrastructure, 13Patch-For-Review: Simplify creating of Jenkins jobs for running browser tests daily - https://phabricator.wikimedia.org/T128190#2169367 (10zeljkofilipin) [15:39:55] (03PS2) 10Paladox: Update npm to 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280912 [15:45:16] hashar: Would updating http://anonscm.debian.org/cgit/pkg-javascript/npm.git/tree/package.json and in engines update npm to 2. work [15:46:36] paladox: no clue [15:46:36] Oh it woulden. You need to update the whole thing [15:46:41] hashar: Oh [15:46:43] at least parsoid repos are all goo [15:46:43] d [15:47:28] hashar: Ok [15:50:15] hashar: https://integration.wikimedia.org/ci/job/npm-node-4.3/3283/console [15:50:22] works for mobilefrontend [15:50:56] hashar: https://gerrit.wikimedia.org/r/#/c/280937/ [15:53:22] (03PS1) 10Paladox: [MobileFrontend] Update npm to npm-node-4.3 [integration/config] - 10https://gerrit.wikimedia.org/r/280940 [15:53:27] hashar ^^ [15:55:04] hashar: COuld this https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/4/console be failing because npm-shrinkwrap needs updating to 200.5 [15:56:30] (03PS1) 10Paladox: [Parsoid] Add npm-node-4.3 test to experimental to test [integration/config] - 10https://gerrit.wikimedia.org/r/280941 [15:56:35] hashar: Could you merge ^^ please. [15:58:23] hashar: How about using this package http://packages.ubuntu.com/xenial/npm which is npm 3 but is made in ubuntu that is based on debian. [16:00:03] well I am off [16:00:05] kids time [16:00:09] might come back later this evening [16:02:43] hashar: Ok [16:15:46] (03CR) 10Dduvall: Provide Rake task to serve as a CI entrypoint (031 comment) [selenium] - 10https://gerrit.wikimedia.org/r/275820 (https://phabricator.wikimedia.org/T128190) (owner: 10Zfilipin) [16:25:55] 10Continuous-Integration-Infrastructure, 6Operations, 6Services: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#2169479 (10Paladox) We can always use http://packages.ubuntu.com/xenial/npm which has a deb package. And since Ubuntu is based on debian it should work. @Hashar and @Krinkle what do... [16:28:11] 10Continuous-Integration-Infrastructure, 6Operations, 6Services: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#2169481 (10Paladox) Or we can use https://github.com/nodejs/node/tree/master/deps/npm and add it in the deps/npm folder. [16:40:04] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2169522 (10greg) p:5Triage>3Unbreak! @Krenair: did you see anything note worthy when looking at this? [16:41:56] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2169535 (10Krenair) When you try to start the instance, Horizon shows a success message in the UI. But when you go to the action log, you find the action listed with mes... [16:53:08] andrewbogott: there's an instance on labvirt1006 that we can't seem to make wake up from horizon or wikitech, could you take a look at the instance? [16:53:17] https://wikitech.wikimedia.org/wiki/Nova_Resource:Deployment-upload.deployment-prep.eqiad.wmflabs [16:59:56] thcipriani: About to go to lunch but I have a few minutes [17:01:15] andrewbogott: deployment-upload in the deployment-prep project on labvirt1006, when you try to start it using wikitech or horizon it shows success, but the logs show 'Error' [17:01:44] it's seemingly permanently SHUTOFF [17:02:09] there's a task for this [17:02:18] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2169622 (10Andrew) nova-compute says: 2016-04-01 17:01:24.328 56558 ERROR oslo_messaging.rpc.dispatcher [req-659dfa94-89f9-41cf-80a4-d81f1a42524a novaadmin deployment-p... [17:02:24] Krenair: ^ that one? [17:02:27] yup [17:02:32] ty [17:03:16] I'll look at the nova source when I have time, might be able to do it this afternoon [17:04:56] andrewbogott: cool thanks :) [17:05:06] 10Continuous-Integration-Infrastructure, 5Gerrit-Migration, 10releng-201516-q2, 3releng-201516-q3, and 2 others: [keyresult] Connect Differential code review with continuous integration - https://phabricator.wikimedia.org/T31#2169636 (10greg) [17:05:14] thcipriani: so, ok hm, i just submitted another patch [17:05:20] i changed the default scap_repository to be false [17:05:24] instead of forcing a convention [17:05:41] This way the default use of scap::source will be just to clone your repo [17:05:56] I also removed the $path parameter, since it looks like scap HAS to work out of /srv/deployment [17:05:58] no matter what [17:06:26] yeah, with the puppet provider, that's definitely a thing [17:07:01] aye [17:07:53] thcipriani: in general, how do you feel about this idea? i know you haven't really gotten to review it yet, but is the idea ok? [17:09:05] thcipriani, I would just bring up a new instance but I think I heard something about that box being unpuppetised [17:09:44] ottomata: this: https://gerrit.wikimedia.org/r/#/c/280771/5/modules/eventlogging/manifests/deployment/keys.pp,cm is waay to complicated, it should be done with hiera [17:09:46] ottomata: This is definitely a thing that needs to happen. I haven't reviewed your patch closely, but I know twentyafterfour has been working on similar stuff. [17:10:22] ottomata: https://gerrit.wikimedia.org/r/#/c/279198/ makes the key configuration waay more straightforward [17:10:40] Krenair: :( it seemingly has a beta role applied, I don't know if that's the sum-total of things that need to happen to rebuild that box. I can give it a look here in a few. [17:10:42] you just override the keys/fingerprints in hiera [17:10:58] ah nice [17:11:02] that is quite nice twentyafterfour [17:11:09] thcipriani: that key stuff existed in a previous patch [17:11:14] in eventlogging::deployment::source [17:11:24] i just moved it into a :;keys class so I could use ::source as a define [17:11:34] * Krenair nods [17:11:54] twentyafterfour: totally happy to move key stuff into hiera and use it from eventlogging::deployment::source [17:12:02] :) [17:12:40] twentyafterfour: i should add you as a reviewer to my change! [17:13:03] HMMM [17:16:28] twentyafterfour: so my patch is mostly dealing with cloning the source repo on the deploy host, and also helping me keep my scap/ configs out of my source repo [17:17:15] i have created a define, scap::source, to do this. but, i suppose this could also be put into hiera instead of in puppet? [17:17:19] via create_resources? [17:17:30] i'm only creating directories and git::clones [17:18:08] https://gerrit.wikimedia.org/r/#/c/280730/8/modules/scap/manifests/source.pp [17:37:01] ottomata: if you want we could try to combine the two patches [17:37:47] hiera definitely makes it cleaner [17:40:17] twentyafterfour: hmmm, let's keep them separate i think, but maybe I could rebase on yours and do things more your way with hiera... [17:40:17] hm [17:40:19] not sure though [17:40:43] i guess it is nice, since all of the scap::sources will go on a single host [17:40:46] (in an env) [17:40:57] its nice to do that via hiera [17:41:06] and then create them all dynamically like you are doing [17:42:39] twentyafterfour: if so, do you think it would be better to keep the scap::source define I made, and create_resource(scap:;source, ...) [17:42:47] or, since scap::source is just a couple of git::clones [17:42:56] create_resource(git::clone, ...) [17:42:56] hmm [17:46:05] (03PS14) 10Dduvall: Provide Rake task to serve as a CI entrypoint [selenium] - 10https://gerrit.wikimedia.org/r/275820 (https://phabricator.wikimedia.org/T128190) (owner: 10Zfilipin) [17:48:25] (03PS15) 10Dduvall: Provide Rake task to serve as a CI entrypoint [selenium] - 10https://gerrit.wikimedia.org/r/275820 (https://phabricator.wikimedia.org/T128190) (owner: 10Zfilipin) [17:50:26] (03CR) 10Dduvall: "Antoine, we've removed the rake_task_spec for now. If you fix it up we can include it in a follow-up patch." (031 comment) [selenium] - 10https://gerrit.wikimedia.org/r/275820 (https://phabricator.wikimedia.org/T128190) (owner: 10Zfilipin) [17:53:24] ostriches: In gerrit 2.12 there is polygerrit preview which is there new gerrit version but has a totally new look. [17:54:03] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2163472 (10thcipriani) hmm...would be nice to recreate that instance; however, the only role applied seems to be `role::beta::uploadservice` which just adds some text to... [17:54:38] https://gerrit.googlesource.com/gerrit/+/master/polygerrit-ui/ [17:56:06] https://groups.google.com/a/chromium.org/forum/#!topic/crashpad-dev/SwYRRr-AT-o [18:05:30] PROBLEM - Host cache-rsync is DOWN: CRITICAL - Host Unreachable (10.68.23.165) [18:06:56] twentyafterfour: Hi i have added a comment here https://secure.phabricator.com/T6878 [18:10:40] (03CR) 10Aashaka: "I agree with the point of reusing Upstream code." [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/279615 (https://phabricator.wikimedia.org/T116561) (owner: 10Aashaka) [18:11:30] ottomata: I'm not sure. create_resource(scap::source) seems right to me [18:11:51] (03PS4) 10Dduvall: Support browser version as part of `BROWSER` [selenium] - 10https://gerrit.wikimedia.org/r/280470 (https://phabricator.wikimedia.org/T128190) [18:12:03] ottomata: since scap::source has the mkdir part in there and you might want to add other stuff later [18:12:21] * twentyafterfour likes scap::source and scap::target [18:13:23] (03CR) 10Aashaka: "Since point 1 and 3 are not according the Mediawiki CC, I feel that we can reuse some of the code, but not just add the complete code at h" [tools/codesniffer] - 10https://gerrit.wikimedia.org/r/279615 (https://phabricator.wikimedia.org/T116561) (owner: 10Aashaka) [18:15:22] twentyafterfour: yaaa ok i think so too [18:15:34] ok then, i'll see if i can amend my patch to make it use hiera like yours does [18:15:37] but most of it will stay the same [18:16:48] twentyafterfour: I never noticed this but you can upload a diff through the web [18:16:49] https://secure.phabricator.com/differential/diff/create/ [18:17:27] (03CR) 10Dduvall: [C: 031] "We removed the rake_task_spec from the previous commit, so this should be good to go now." [selenium] - 10https://gerrit.wikimedia.org/r/280470 (https://phabricator.wikimedia.org/T128190) (owner: 10Dduvall) [18:31:09] twentyafterfour: We can use the diff tool from the web see my patch using it https://secure.phabricator.com/D15563 [18:31:41] I used my normal install but instead of doing git push i did git show > patch.txt and then uploaded the patch to the diff on the web [18:31:50] greg-g ^^ [18:32:14] greg-g we can use phabricator diff tool. I found that really easy. And for now a woraround to not use arc. [18:35:02] There is even `arc diff --preview`, a diff not attached to a revision. And you can also have a revision without repo, for example to already review the code of a new component pending repo creation, or to report upstream. [18:35:18] 5Gerrit-Migration, 10Differential: Find way to use Differential with plain git (i.e.: without requiring arc) - https://phabricator.wikimedia.org/T127#2169997 (10Paladox) @greg and @mmodell we can use phabricator diff on the web tool. See https://secure.phabricator.com/differential/diff/create/ We should enab... [18:36:08] ah the goal is not to use arc, okay [18:36:47] Dereckson: Yeh, but it is still recommended to use it for now. But there are others ways as described above that users doint have to use arc. [18:36:50] twentyafterfour: do you mind if I amend your patch? i want to change the deploy_host class name, and also fix tabbing :) [18:37:45] Dereckson: It is already enabled at https://phabricator.wikimedia.org/differential/diff/create/ much easy now to upload patches. [18:38:16] greg-g and twentyafterfour: Maybe we can describ https://phabricator.wikimedia.org/differential/diff/create/ as an alternitive thing. [18:38:39] Plus it will allow other users using the gerrit tool that uploads patches to gerrit to be able to switch. [18:39:00] hmmm sorry, I even maintain a Docker image of arc to use it on a server we build Docker images on, not really interested by any willingness to skip it. [18:40:37] Dereckson: Not sure what you mean. [18:41:00] We are trying to allow users to use there normal git. But they can also use arc if they want. [18:42:30] paladox: yeah I knew about https://secure.phabricator.com/differential/diff/create/ [18:42:55] paladox: I'm not sure why a web form is better than arc though [18:43:04] ottomata: I don't mind [18:43:04] twentyafterfour: I find using that much easy since it dosen't seem anytime soon that using git push will be supported. [18:43:17] ottomata: what tabbing should I have used? [18:43:21] twentyafterfour: Just that arc is new to windows and is hard to setup. [18:43:38] https://phabricator.wikimedia.org/D168 [18:43:41] twentyafterfour: puppet is 4 space [18:43:45] you got 2 in deploy_host [18:43:52] paladox: I see. I never use windows myself but I don't think it'd be that hard to set up if you already have git working on windows [18:44:24] ottomata: hmm I see other places where 2 was used, I thought that was the norm for puppet [18:45:03] twentyafterfour: It was when i was doing it for phabricator. Kept saying something that there is an error in the lint. And wasen't todo with the patch i uploaded. It took me a couple of hours to find the command that force pushes and by passes lint but i forgot now. [18:45:04] * twentyafterfour is indifferent to tab size but the 80 char line length limit encourages me to use smaller tabs [18:45:53] twentyafterfour: i prefer 2 spaces too, but years ago ops/puppet tab policy was settled :/ [18:45:54] paladox: arc diff --nolint [18:46:15] twentyafterfour: Oh ok. [18:46:18] its two spaces in yaml though :/ [18:46:44] ottomata: that must be what got me confused [18:46:45] 5Continuous-Integration-Scaling, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate javascript npm CI jobs to Nodepool - https://phabricator.wikimedia.org/T119143#2170061 (10hashar) [18:46:47] 5Continuous-Integration-Scaling, 10OOjs, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate OOjs npm CI job to Nodepool - https://phabricator.wikimedia.org/T128089#2170058 (10hashar) 5Open>3Resolved a:3Paladox Somehow composer is not/no more needed. [18:47:14] 5Continuous-Integration-Scaling, 10OOjs-UI, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate OOjs UI npm CI job to Nodepool - https://phabricator.wikimedia.org/T128091#2063634 (10hashar) Depends on composer being installed. [18:48:41] twentyafterfour: Is there a way to get jenkins to show harbormaster-test as the text but in green and then that should link to the build instead of showing a blue link. [18:48:45] https://phabricator.wikimedia.org/D115 [18:48:46] 10Continuous-Integration-Infrastructure, 7Regression: Jobs sometimes fail with "/usr/local/bin/npm: No such file or directory" - https://phabricator.wikimedia.org/T129617#2170064 (10Krinkle) 5Open>3Resolved a:3Krinkle I've removed `/usr/local/bin/npm` from those slaves with salt so that it's now consiste... [18:49:21] paladox: the thing is, arc helps you format your commit messages, it figures out the base revision so you don't have to do a bunch of git wrangling, it helps with amending and squashing patches when you want to update a diff or land an accepted diff... so arc provides a lot of helpful benefits. Plus we are looking at standardizing all of our unit testing with `arc unit` [18:50:17] 5Continuous-Integration-Scaling, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate javascript npm CI jobs to Nodepool - https://phabricator.wikimedia.org/T119143#2170068 (10hashar) **Status update** Most repositories have been migrated. What is left: | Task | Repo | Comment |--|--|-- | T130207 | pywi... [18:50:23] so you miss out on a lot of benefits. Hopefully arc will get better on windows, especially with windows 10 supporting a unix environment (https://insights.ubuntu.com/2016/03/30/ubuntu-on-windows-the-ubuntu-userspace-for-windows-developers/) [18:50:32] twentyafterfour: Oh but what does arc unit do. Does that mean we wont need phpunit, jshint and jscs any more. Even though we use them. [18:51:13] twentyafterfour: Yeh, i herd ubuntu is being built in windows. It is already some what supported since someone found a linux script or linux file in windows. [18:51:31] I use windows insiders so hopefully soon we can to test make and the native bash. [18:52:44] 10Continuous-Integration-Config, 5Continuous-Integration-Scaling, 10releng-201516-q3, 7WorkType-NewFunctionality: [keyresult] Migrate php composer (Zend and HHVM) CI jobs to Nodepool - https://phabricator.wikimedia.org/T119139#2170074 (10hashar) **Status update** Prerequisite tasks have been accomplished... [18:53:55] twentyafterfour: We get access to apt too. [18:55:13] twentyafterfour: Looks like apache is being installed [18:55:14] :) [18:55:23] paladox: arc unit is just a wrapper that calls various test frameworks, it doesn't replace them entirely [18:55:37] twentyafterfour: Oh ok [18:57:18] arc lint is more useful, as it calls successive linters [18:58:28] Dereckson oh ok, i may try it later. [18:58:53] I get access to ssh script now :) no longer need putty [19:01:51] twentyafterfour: Looks like we will be getting access to 16.04 LTS in the windows store soon. [19:05:04] ta dah! [19:05:07] heya [19:05:08] ok so [19:05:08] * apergos waits for applause [19:05:12] clap clap clap [19:05:13] * Reedy claps [19:05:17] * apergos takes a bow [19:05:18] ok [19:05:21] o [19:05:22] so [19:05:24] this is the earth [19:05:27] it's rooouund [19:05:30] flat [19:05:36] in your heart, you know it's flat [19:06:11] 10Continuous-Integration-Infrastructure, 6Labs, 6Operations, 10Packaging: Update phantomjs to 2.1.1 on trusty - https://phabricator.wikimedia.org/T130940#2170134 (10Krinkle) 5Open>3declined There is no need for this. CI packages should be installed locally on a per-needed basis. This is just the need o... [19:06:28] so I understand that the deployment user that does the scap deploy on tin/mira should own or have rights to write in the tin/mira copy of the repo [19:06:33] for .git at least, yes? [19:07:40] apergos: no isn't necessary [19:07:47] no? [19:07:50] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2163472 (10hashar) deployment-upload is barely puppetized. It has been hacked in a few years ago as a way to emulate the media servers we have used before migrating to... [19:08:00] PROBLEM - App Server Main HTTP Response on deployment-mediawiki03 is CRITICAL: HTTP CRITICAL: HTTP/1.1 500 Internal Server Error - 2069 bytes in 0.105 second response time [19:08:08] are you asking about doing things like git pulls where you don't own .git? [19:08:16] PROBLEM - App Server Main HTTP Response on deployment-mediawiki01 is CRITICAL: HTTP CRITICAL: HTTP/1.1 500 Internal Server Error - 2068 bytes in 0.072 second response time [19:08:23] about the whole process [19:08:24] PROBLEM - English Wikipedia Main page on beta-cluster is CRITICAL: HTTP CRITICAL: HTTP/1.1 500 Internal Server Error - 2470 bytes in 0.078 second response time [19:08:28] PROBLEM - App Server Main HTTP Response on deployment-mediawiki02 is CRITICAL: HTTP CRITICAL: HTTP/1.1 500 Internal Server Error - 2068 bytes in 0.075 second response time [19:08:32] PROBLEM - English Wikipedia Mobile Main page on beta-cluster is CRITICAL: HTTP CRITICAL: HTTP/1.1 500 Internal Server Error - 2470 bytes in 0.099 second response time [19:08:45] I don't actually care if it has to be root but I assume that we don't like that [19:08:52] and we prefer an actual regular user account [19:09:01] ja [19:09:02] to manually run scap from tin/mira, yes? [19:09:03] ok [19:09:06] so, on tin [19:09:08] https://git-scm.com/docs/git-init [19:09:09] see --shared [19:09:13] i'm pretty sure that's how that works! [19:09:17] maybe 90% sure [19:09:19] but basically [19:09:24] so in my case I would tell people "if you can be the datasets user, then you can deploy" but this would entail [19:09:29] having that user on tin/mira [19:09:29] the git clone there is group owned and writeable by wikidev [19:09:32] how do I make that happen [19:09:36] oh but see [19:09:38] that's wikidev [19:09:47] that's not applicable for the dumps repo [19:09:47] and the dir has a umask of 002 [19:09:52] why? [19:09:55] this is just on the deploy host [19:09:57] not on the taret [19:09:58] target [19:10:10] how do things get changed on the target side? [19:10:14] on target side [19:10:19] there are ssh keys set up for a special deploy user [19:10:25] that's the deploy_user you set with scap::target [19:10:37] so here's the bit where I don't knoqw whose keys those ought to be then [19:10:41] you have to be in a certain group on tin in order to access the ssh keys [19:10:42] 10Continuous-Integration-Infrastructure: Deal with bitbucket.org rate limit for downloading phantomjs (npm jobs intermittently fail) - https://phabricator.wikimedia.org/T127100#2170156 (10Krinkle) After 2 weeks of extremely crippled productivity in OOjs UI due to Jenkins failing 9/10 times, @Jdforrester-WMF and... [19:11:02] the, when you do a scap deploy [19:11:11] you ssh in as the deploy_user using those keys [19:11:25] and a git fetch and checkout is run on the target [19:11:27] as that deploy_user [19:11:31] from tin [19:11:34] ahh I see [19:11:44] 10Continuous-Integration-Infrastructure: Deal with bitbucket.org rate limit for downloading phantomjs (npm jobs intermittently fail) - https://phabricator.wikimedia.org/T127100#2170163 (10Paladox) Well they updated https://www.npmjs.com/package/phantomjs-prebuilt to 2.1.7 which they now host phantomjs binarys on... [19:11:56] so here is where the dumps_deploy key can belong to the datasets user which will already be defined on the targets [19:12:25] Krinkle hi, see https://github.com/Medium/phantomjs please. They now have the binarys on github. [19:12:25] all right, I think that holds me for now [19:12:25] ja you can set deploy_user => 'datasets' when you decalre the scap::target [19:12:31] yes I have that already [19:12:39] but clearly I didn't understand what exactly it did :-D [19:12:47] i'm a little hazy on how the key gets created and sahred, but i think its just like this [19:12:51] https://gerrit.wikimedia.org/r/#/c/279198/10/hieradata/role/common/deployment/server.yaml [19:12:55] I thought there was something to be set at both ends but it turns out that trustedgroup thing is all [19:13:11] so, people in the trusted_group can access the key [19:13:15] on the deploy host [19:13:19] yeah I think I just create it on palladium and stuff the private key into secrets, put the pub key into keyholder somehow or other [19:13:26] declare em and go [19:13:31] well, the scap::target does make you set the public key [19:13:39] so the public key ends up on the target [19:13:46] right, one gets the fingerprint and the other the key [19:13:48] right [19:14:06] I think that's allt he pieces I need for right now except it'as obvious I won't try to set this up tonight [19:14:12] I'll read the patchs and [19:14:17] is there a task for this work you're doing? [19:14:22] I'd subscribe :-P [19:14:27] well, twentyafterfour is doing the keyholder thing [19:14:29] which i just leared about today! [19:14:34] which is awesome [19:14:43] and so is the scap directory stuff [19:14:43] i was doing work yesterday on bootstrapping for scap sources [19:14:50] those two pieces are what I really want [19:14:52] which thus far hadn't been done [19:15:06] I was looking at the eventbus repo on tin [19:15:08] i started this work for [19:15:08] https://phabricator.wikimedia.org/T118772 [19:15:19] and was scared to see that there was a big fat notice in the scap dir saying [19:15:23] haha [19:15:23] 'yeah these files aren't in a repo" [19:15:25] yup :/ [19:15:26] paladox: Yeah, we can use that next time there is an update. [19:15:28] hence this work! [19:15:29] I think it as eventbus anyways [19:15:32] so [19:15:36] Right now it's just the same version (2.1.1) so there's no point in changing everything back now. [19:15:36] since i want to move eventlogging to use scap too [19:15:39] I want to avoid that :-D [19:15:41] but i don't want to put scap in the source repo [19:15:42] exactly! [19:15:46] EXACTLY [19:16:00] u hearin this thcipriani and twentyafterfour everybody wants it :p [19:16:00] subscribed [19:16:14] you can use scap to sync the scap version inside mediawiki and then use the scapped scap to scap your scap onto the app servers. [19:16:20] let me put it this way, I will bootstrap that sucker some ugle way if I don't get a nice way [19:16:22] i don't see a phab ticket associated with twentyafterfour's keyholder patch [19:16:24] but it WILL be in puppet [19:16:29] hm [19:16:30] Krinkle ok, ive updated it here https://github.com/jdforrester/grunt-svg2png/pull/15 and https://gerrit.wikimedia.org/r/#/c/280781/ [19:16:35] maybe I'll make him one :P [19:18:05] RECOVERY - App Server Main HTTP Response on deployment-mediawiki03 is OK: HTTP OK: HTTP/1.1 200 OK - 40229 bytes in 6.486 second response time [19:18:15] RECOVERY - App Server Main HTTP Response on deployment-mediawiki01 is OK: HTTP OK: HTTP/1.1 200 OK - 40212 bytes in 0.615 second response time [19:18:23] RECOVERY - English Wikipedia Main page on beta-cluster is OK: HTTP OK: HTTP/1.1 200 OK - 40559 bytes in 0.659 second response time [19:18:29] RECOVERY - App Server Main HTTP Response on deployment-mediawiki02 is OK: HTTP OK: HTTP/1.1 200 OK - 40218 bytes in 0.625 second response time [19:18:31] RECOVERY - English Wikipedia Mobile Main page on beta-cluster is OK: HTTP OK: HTTP/1.1 200 OK - 29736 bytes in 0.637 second response time [19:19:21] 10Continuous-Integration-Infrastructure: Deal with bitbucket.org rate limit for downloading phantomjs (npm jobs intermittently fail) - https://phabricator.wikimedia.org/T127100#2032437 (10hashar) > After 2 weeks of extremely crippled productivity in OOjs UI due to Jenkins failing 9/10 times To be fair, that was... [19:20:31] found it and added the changeset [19:20:39] https://phabricator.wikimedia.org/T130419 [19:21:14] 5Gerrit-Migration, 10Differential: Find way to use Differential with plain git (i.e.: without requiring arc) - https://phabricator.wikimedia.org/T127#2170194 (10greg) >>! In T127#2169997, @Paladox wrote: > @greg and @mmodell we can use phabricator diff on the web tool. > > See https://secure.phabricator.com/d... [19:24:52] (03CR) 10Hashar: [Parsoid] Add npm-node-4.3 test to experimental to test (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/280941 (owner: 10Paladox) [19:25:03] hashar: Hi [19:26:01] (03PS2) 10Hashar: [Parsoid] Add npm-node-4.3 test to experimental to test [integration/config] - 10https://gerrit.wikimedia.org/r/280941 (owner: 10Paladox) [19:26:53] (03CR) 10Paladox: "Thanks." [integration/config] - 10https://gerrit.wikimedia.org/r/280941 (owner: 10Paladox) [19:26:55] (03CR) 10Hashar: [C: 032] [Parsoid] Add npm-node-4.3 test to experimental to test [integration/config] - 10https://gerrit.wikimedia.org/r/280941 (owner: 10Paladox) [19:27:17] ottomata: yeah there isn't a ticket for that I guess, I should have made one [19:27:32] oh I guess there was [19:27:45] * twentyafterfour is forgetful sometimes [19:27:57] (03Merged) 10jenkins-bot: [Parsoid] Add npm-node-4.3 test to experimental to test [integration/config] - 10https://gerrit.wikimedia.org/r/280941 (owner: 10Paladox) [19:28:14] twentyafterfour: good morning! Looks like the Phabricator server can talk to Gearman on Gallium/Zuul just fine can't it? [19:28:35] for some reason I thought it needed tweak of some other hidden firewall [19:28:54] hashar: because the fix already got merged? [19:29:56] hashar: https://phabricator.wikimedia.org/T131375#2166726 [19:30:58] hashar: Could we test running jenkins on one of the repos in phabricator. I can upload patches through the diff tool on the phabricator.wikimedia.org [19:32:34] hashar: and twentyafterfour: What about using this jenkins plugin https://wiki.jenkins-ci.org/display/JENKINS/Phabricator+Differential+Plugin for differential. [19:32:40] https://wiki.jenkins-ci.org/display/JENKINS/Phabricator+Differential+Plugin [19:32:44] https://github.com/uber/phabricator-jenkins-plugin [19:34:28] paladox: we already use that [19:34:41] twentyafterfour: oh ok. [19:34:53] paladox: we use differential for scap development already and it has tests in jenkins [19:35:20] twentyafterfour: yeah somehow I thought it would have been much more complicated :D [19:35:31] 10Continuous-Integration-Infrastructure: Deal with bitbucket.org rate limit for downloading phantomjs (npm jobs intermittently fail) - https://phabricator.wikimedia.org/T127100#2170242 (10Krinkle) 5Open>3Resolved a:3Krinkle >>! In T127100#2170183, @hashar wrote:> > Maybe we can have a mirror on wmflabs?... [19:35:34] twentyafterfour: I tested with https://phabricator.wikimedia.org/D169 but tests failed because i did not use arc unit tests. [19:36:16] paladox: e.g. https://phabricator.wikimedia.org/D150 see the tests work with jenkins [19:36:31] PROBLEM - Free space - all mounts on deployment-sentry2 is CRITICAL: CRITICAL: deployment-prep.deployment-sentry2.diskspace._var.byte_percentfree (<10.00%) [19:36:36] twentyafterfour: Yep. [19:36:53] paladox: your tests failed because you didn't use arc :P [19:38:10] twentyafterfour: Yes, when we start migrating repos such as extensions like CentralAuth and we upload a patch to differential using the diff tool i used and it runs the jenkins tests. Will they be tested the same so same tests. And not needing arc to upload the patch. [19:38:49] twentyafterfour: Even if i uploaded the patch with arc its because i did not run the arc unit tests so with windows they fail so i would need to ignore them. [19:39:39] twentyafterfour: Would running a recheck command there re run the tests. [19:39:53] hmm, hey twentyafterfour i'm testing your patch in deployment prep [19:40:08] Could not find resource 'Group[deploy-phabricator]' for relationship on 'Keyholder::Private_key[phabricator_rsa] [19:40:13] paladox: do you want an account on a FreeBSD server where there are git, git-review, arc, phpunit, npm, etc. already installed? [19:40:17] paladox: currently you need to use arc to run tests [19:40:28] i think you might not want to add the dependency here Group[$trusted_group] -> Keyholder::Private_Key[$key_file] [19:40:28] so you wouldn't be stuck on Windows. [19:40:31] in keyholder::agent [19:40:39] especially for data.yaml managed groups [19:40:40] Dereckson: Oh, yes please. [19:40:51] twentyafterfour: Ok [19:40:59] paladox: we are still working on it, give us time [19:41:02] Okay. [19:41:03] Ok [19:41:13] ottomata: ok why not add the dependency? [19:41:22] hm, i'm not sure if data.yaml works that way [19:41:31] i don't think it declared puppet group resources that you can depend on [19:41:35] or..am I wrong? [19:41:38] (03PS2) 10Hashar: [MobileFrontend] Update npm to npm-node-4.3 [integration/config] - 10https://gerrit.wikimedia.org/r/280940 (owner: 10Paladox) [19:41:41] declares* [19:41:55] geez puppet-compiler causes my browser to freeze every time I open the results page [19:42:02] (03CR) 10Hashar: "Marked with depends-on." [integration/config] - 10https://gerrit.wikimedia.org/r/280940 (owner: 10Paladox) [19:42:10] ottomata: in the end it does [19:42:39] ottomata: I think anyway, it's kinda hard to follow the chain of includes / resources / abstractions ... but eventually it creates admin::groups or something [19:42:56] hmmmm maybe it does [19:43:05] hm you might be right, i'm reading that now too [19:43:14] hmmm, maybe the grou pjust doesn't exist on deployment-tin [19:43:31] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 60.00% of data above the critical threshold [0.0] [19:43:36] ottomata: that's probably the case, I had the same problem with phabricator [19:43:36] yeahh that must be it [19:43:47] I had to add the group on tin [19:44:09] hm, twentyafterfour, probably it should just use a different group, ja? [19:44:16] in labs? [19:44:21] like i was doing for eventlogging [19:44:38] - "project-${::labsproject}" [19:45:07] hashar: There's polygerrit in gerrit 2.12 [19:45:24] It has a better ui then gerrit 2.12 has but still needs alot of work. [19:45:31] https://gerrit.googlesource.com/gerrit/+/master/polygerrit-ui/ [19:45:41] ottomata: that might work [19:46:01] I hate having things differ between production and beta though - it always leads to surprises [19:46:02] hm, maybe twentyafterfour, we should make a default for $trusted_group in keyholder agent [19:46:06] paladox: I dont think we will advertise polygerrit [19:46:18] paladox: since we go for Differential [19:46:21] ottomata: a default for beta or a global default [19:46:23] ja, but you should think about it as more than beta, any puppet you design should work in any labs project [19:46:33] twentyafterfour: both [19:46:35] i'd do [19:46:38] hashar: Ok [19:46:53] in labs default to $::labsproject [19:46:56] else uhh [19:47:02] default to i dunno... deploy-service? [19:47:13] HMM, no [19:47:18] ottomata: with hiera you can override the vars in wikitech hiera interface [19:47:24] ja [19:47:24] true [19:47:33] this won't work, because hiera_hash will pull in the common,yaml [19:47:34] first [19:47:43] that's how I intended to handle different wikitech projects using it [19:47:43] so, the default in the class won't be respected anyway [19:47:45] aye [19:47:48] hashar: Would we be able to test some of the tests we use on jenkins on differental. So tests like npm and extension-unittests [19:47:49] please [19:47:52] ok, so either way then [19:48:13] I intend to use this for the phabricator wikitech project, so definitely not just beta [19:48:22] aye [19:48:23] ok, um [19:48:35] shoudl we change the trusted_group fro deployment-prep, or are you going to ensure the group exists? [19:48:55] not entirely sure how to do that in labs.. [19:48:58] in fact it's phabricator deplotments that lead me to do all of this hiera stuff because I have so many things that need to be overridded (phab deploys are different from other services in several unfortunate ways)\ [19:49:09] hmm [19:49:10] hieradata/labs/deployment-prep/host [19:49:11] ? [19:49:27] there is a deployment-tin there [19:49:33] ottomata: I'm not entirely sure how hieradata is evaluated [19:49:43] yeah host would work [19:49:52] hashar: In windows i will soon get to use bash natavly and ssh commands plus php and apache commands plus apt. and apt update. Since they are building ubuntu inside of windows. I will get to use ubuntu 16.04 when it hits the windows store [19:49:53] or we can make a deployment role of some sort [19:50:13] mira.deployment-prep exists but isn't currently used so it's just tin really in practice [19:50:13] admin::groups [19:50:34] I wish we could duplicate all of the production groups [19:50:42] yeah. [19:50:43] so that it wouldn't need separate config [19:50:49] twentyafterfour: maybe you should just do labsproject for this [19:50:55] since, in beta, everyone is a member of that [19:50:59] paladox: do you mean Microsoft is going to ship a Ubuntu virtual machine for Windows? [19:51:00] so anyone then will be able to deploy [19:51:03] yeah that works [19:51:07] ok [19:51:09] amending... [19:51:13] ottomata: that sounds good [19:51:47] the only problem is it hides some complexity and then when you move to production you are hit with a new thing you didn't think about (the groups) [19:52:04] oh HM [19:52:08] twentyafterfour: trusted_group can be an array? [19:52:23] hashar: No there actually building ubuntu into windows. It is not virtual and is not a container. See https://insights.ubuntu.com/2016/03/30/ubuntu-on-windows-the-ubuntu-userspace-for-windows-developers/ which was linked by twentyafterfour [19:52:39] i think that will break Group[$trusted_group] -> Keyholder::Private_Key[$key_file] [19:52:59] hashar: I can also run ubuntu packages to inside of windows [19:54:04] oh my god [19:54:31] ottomata: yes [19:54:46] hm, maybe it won't? [19:54:48] hmm good point [19:54:51] I don't know for sure [19:54:52] http://serverfault.com/questions/616116/how-can-i-require-an-array-of-resources-in-puppet [19:55:00] that says you can do it! in the only answer [19:55:12] ok let's try it i guess [19:55:45] paladox: that is clever really. When I was running Windows I used gnuwin32 ( https://en.wikipedia.org/wiki/GnuWin32 ) i.e. native ports to Window [19:56:18] hashar: Oh, do you still run windows or not any more. [19:56:23] paladox: good to see Microsoft/Ubuntu officially endorsing that. My next computer might well be a Windows based machine, this way I can play games again hehe [19:57:06] hashar: Yeh, windows 10 is shapping to be good. But only if microsoft sort out there privacy and not ready what everyone types in the keybored. [19:57:12] paladox: been using Linux/Windows since middle 90's til roughly 2007-2008 when I switched to Mac (and still used Window a bit for gaming) [19:57:28] but since ~2010 pure mac, and a bit of linux from time to time [19:57:56] They have 7m windows insiders, i may get an apple mac, ive been using the ipad pro for a few months very powerful but they doint allow you to do what you can on an apple mac. [19:57:59] paladox: I suggest just installing linux ;) [19:58:34] I have zero regrets from abandoning windows entirely... windows xp was the last version I ran, about 8 years ago [19:58:49] HMMM oh duhh twentyafterfour, there's no group { project-$labsproject in puppet either [19:58:50] same problem [19:58:55] ERrrrrrrr [19:59:01] dunno about this dependencyYyYyyy [19:59:02] twentyafterfour: Oh, i doint really like linux, i used it in a virtual program. [19:59:03] yes it *should* be there [19:59:04] hm [19:59:15] ottomata: it's ok if you want to remove the dependency [19:59:35] yeah, i think we might have to, the comments to keyholder::agent do say that someone else is repsonsible for ensuring trusted_group exists [19:59:50] hashar: You can run mutiple virtual windows inside windows. Meaning you can have microsoft edge open in one virtual window and chrome in the other. [19:59:59] paladox: there are tons of different desktop configs, you might not like unity but you might like xfce or kde or gnome... it's a lot of shit to figure out but once you find the right setup then it's awesome [20:00:14] Ok [20:00:20] wich are all equally crap compared to Mac OS X interface ;-D [20:00:24] [20:00:24] ottomata: ok [20:00:55] Yeh, i thought mac is run of open source unix but then closed sourced it. [20:02:06] hashar: I can also run grep now in windows, that is once they release a windows 10 build with ubuntu. [20:03:04] Looks like they are teasing something at https://twitter.com/GabeAul/status/715987332814364673 hoping a new build comes out, its been like a week since last release but that was more stability then features. [20:04:09] twentyafterfour: I can run the commands that are on ubuntu soon in windows. [20:04:54] yes [20:05:00] hashar: You can run windows in apple bootcamp. [20:05:23] paladox: who'd want to run that? :P [20:05:56] twentyafterfour: do you mean apple bootcamp or ubuntu in windows [20:06:09] You can run Windows in VirtualBox much easier (unless you're playing high-level graphics games and need the full CPU) [20:06:13] (on Mac) [20:06:15] just windows ... /me hates windows [20:06:34] yeah running windows outside of a VM is something I will stay away from [20:06:39] * twentyafterfour doesn't trust it [20:06:44] s/outside of a VM// [20:07:05] but I do have that privilege, since I don't have to develop and test for browsers [20:07:55] twentyafterfour: Yeh running windows 7 in a virtual machine was really slow. But lucky i have windows 10 now. Much better plus they brought back the start button. [20:09:13] apergos: I worked actively on jQuery and not run windows on my machine (VM or otherwise). Using a local VM for broswer testing is so 2008. [20:09:17] Krinkle: do you know whether we still need grunt-cli to be installed by the puppet contint::javascript class ? [20:09:39] hashar: They price the apple macs too much. I wish apple would give you more storage for less money. Since a 1tb hard drive with windows is £299 pound here. But i found 2tb for 30 pounds more [20:09:43] apergos: yeah I don't run windows in a vm either, anymore. but I used to do a lot of web dev and at one time I even did visual studio development so I kinda had to [20:10:26] Krinkle: 2008 was the last time I ran windows. lol [20:10:32] hashar: I don't know. You should check https://github.com/search?l=json&q=%22test+grunt%22+%40wikimedia&ref=searchresults&type=Code&utf8=%E2%9C%93 and check if any of them don't specify grunt-cli [20:10:32] something like that anyway [20:10:42] oh [20:10:56] (in devDependencies) [20:11:18] we got the mediawiki extensions migrated to Nodepool instances which do not have grunt-cli -D [20:11:38] Yes, using a 2 year old, EOL, unmaintained, insecure, buggy and slow version of npm. [20:11:57] I'm gonna have to require testing on Travis CI before any important releaes. [20:11:57] blame services for upgrading node without backporting npm ? :D [20:12:07] I am going to provision it using npm install [20:12:15] Services don't need npm. Be thankful that they packaged node for us. [20:12:42] Or really blame Ubuntu for breaking node in the first place. [20:12:58] (and then not caring to update it ever) [20:13:26] RECOVERY - Puppet run on deployment-tin is OK: OK: Less than 1.00% above the threshold [0.0] [20:13:51] Krinkle looks like ubuntu will get alot of traffic from 200 milion plus users or only the ones that like coding. From windows 10. Since we will hopefully soon get to install packages [20:14:45] paladox: You mean windows 10 will be able to install aptitude/debian packages from ubuntu? [20:14:49] yes [20:14:58] I thought they created their own package manager that looks like the one from ubuntu. [20:15:01] It does not use ubuntu. [20:15:12] and that wouldn't work anyway, unless windows suddenly became posix. [20:15:20] Krinkle: we got Node 4.3 from Debian, but Debian does not have bumped npm (it is a split package) [20:15:22] Krinkle microsoft built something with the help of ubuntu that translates linux code into windows [20:16:28] Krinkle see [20:16:29] https://insights.ubuntu.com/2016/03/30/ubuntu-on-windows-the-ubuntu-userspace-for-windows-developers/ which was linked by twentyafterfour [20:16:32] please [20:18:42] Interesting. [20:19:10] Yep, i get bash natavaly now. [20:19:33] Hello, can someone add me? I want to make some querys at the beat cluster for testing purposes. (Read-only would be enough) [20:20:05] Krinkle: Also looking for users in windows on bash it would look like /mnt/User/ but on git for windows it would be c:\users [20:21:17] An easier solution would've been for Windows to add some emulation bindings the other way around. You know, ditch windows and let Microsoft acquire WineHQ. [20:21:39] apergos: :P [20:21:54] heh [20:22:52] But yeah, exciting progress. [20:23:57] Krinkle also they say the performance is the same using it as if using it on ubuntu [20:23:59] (03PS1) 10Hashar: dib: provision npm 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280980 (https://phabricator.wikimedia.org/T124474) [20:24:16] 22:19 < Luke081515> Hello, can someone add me? I want to make some querys at the beat cluster for testing purposes. (Read-only would be enough) [20:24:22] can someone help me with that? [20:25:05] twentyafterfour: i think maybe you meant to include scap::server (previously scap::deploy_host)in role::deployment::server [20:25:06] yes? [20:25:07] not sure. [20:25:08] Krinkle: and Iam dropping grunt-cli via https://gerrit.wikimedia.org/r/#/c/280974/1/modules/contint/manifests/packages/javascript.pp,cm [20:25:11] you put the yaml in that role [20:25:21] you included it in scap::master [20:25:28] which is included from role::deployment::mediawiki [20:25:28] hashar: k. we should probably announce it just in case. [20:25:33] yup [20:25:45] hashar: OT: Are there criterias for read-only access to the db at beta? [20:26:52] Krinkle: announcing to wikitech / QA [20:27:12] Luke081515: I dont think we have any policy regarding adding user to the beta cluster project [20:27:46] Luke081515: anyone getting shell access would get write access to the DB which does not hold that many data but does have some private info just like any project in labs [20:27:50] so it is probably free form [20:28:17] hashar: Ok, can you add me? This would support my developments [20:31:49] !log Dropping grunt-cli from the permanent slaves. People can have it installed by listing it in their package.json devDependencies https://gerrit.wikimedia.org/r/#/c/280974/ [20:31:54] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [20:37:24] !log Added Luke081515 as a member of deployment-prep (beta cluster) labs project [20:37:29] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [20:37:36] hashar: Thank you very much [20:39:36] greetings Luke081515 :) [20:39:45] Hi FlorianSW :) [20:40:04] Luke081515: so in short [20:40:28] Luke081515: dont change random stuff, the instances are configured solely via puppet / hiera config siwtches [20:40:38] Luke081515: running random commands WILL break something [20:40:39] * Luke081515 will be careful [20:40:54] so better ask twice here on IRC :-} [20:41:07] my goal is not to destroyinfrasturcture from other users :) [20:41:15] and if you ever screw up something somehow, no big deal we all make mistakes but please: A) !log the mistake [20:41:24] ok [20:41:29] B) eventually fill a task giving as many detail as possible so we get clue as to what got broken [20:41:34] again, mistake is acceptable [20:41:41] hiding the mistake is unacceptable ;-} [20:42:03] ok :) [20:42:16] I would ban folks for hiding stuff. But will work hard to get an issue fixed if it is disclosed :-} [20:42:18] Yippee, build fixed! [20:42:19] Project selenium-Echo » chrome,beta,Linux,,contintLabsSlave && UbuntuTrusty build #12: 09FIXED in 1 min 18 sec: https://integration.wikimedia.org/ci/job/selenium-Echo/BROWSER=chrome,MEDIAWIKI_ENVIRONMENT=beta,PLATFORM=Linux,VERSION=,label=contintLabsSlave%20&&%20UbuntuTrusty/12/ [20:42:37] hashar: First I plan to use only read, so I guess I waon't break something [20:42:44] the main work machine is usually deployment-tin.deployment-prep.eqiad.wmflabs [20:43:02] you should get access to the DB using: sql enwiki [20:43:13] which should yields a prompt like: (wikiadmin@deployment-db2) [enwiki]> [20:43:18] but some time again I broke a lot of at beta, because I used my read access onwiki :D. A flow query got an error [20:43:27] and to get write access: sql --write enwiki [20:43:31] again do abuse the !log command :-} [20:43:36] ok :D [20:43:47] and labs terms policy apply [20:43:59] so dont copy paste password / IP / emails etc [20:44:07] should not even get on your machine -:} [20:44:17] but that I bet is common sense for anyone familiar with wmflabs [20:44:22] otherwise, you should be fine [20:44:38] yeah, I got already one project [20:44:46] So I'm prefering to vandalise there :D [20:44:47] http://logstash-beta.wmflabs.org/ usually has the good error logs, I think we even get syslog messages relayed there [20:44:57] ;-} [20:45:07] ok, thank you [20:45:15] hashar: Could you remove npm verbose or the hack we did earler from today on https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/6/consoleFull [20:45:26] Seems to fill and almost crash my browser. [20:45:31] please [20:45:41] hashar: Please beware that the puppet class may need changing [20:45:48] https://integration.wikimedia.org/ci/job/npm-node-4.3/3318/console [20:45:55] npmrc on ubuntu is not the same as on jessie [20:46:04] the jessie defaults to prefix=/usr/local apparently [20:46:08] which is good I guess [20:46:40] Krinkle: apparently it installs to /usr/bin/npm [20:46:50] oh [20:46:54] that is the nodepool instance [20:46:59] Yes [20:47:00] Jessie [20:47:14] paladox: btw, I'm glad you're helping with the arc/differential stuff, I just want to keep our production instance focused on real work, we have https://phab-01.wmflabs.org/ if you want to mess around (I've done a bit, eg: https://phab-01.wmflabs.org/tag/project_-_greg_grossmeier/ :) ) [20:47:30] paladox: ^^ that's re that test change you sent up :) [20:47:55] greg-g: Ok, thanks [20:48:05] Krinkle: so the npm Debian package has prefix = "/usr/" [20:48:30] err forget me [20:48:33] Krinkle: so the npm Debian package has prefix = "/usr/local" [20:48:38] Krinkle: but if you npm install -g npm@2.4.x using the puppet class, the prefix is "/usr" [20:48:43] hashar@integration-slave-jessie-1001:~$ npm conf ls -l|grep prefix [20:48:43] prefix = "/usr" [20:48:57] somehow [20:51:27] 5Continuous-Integration-Scaling, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate javascript npm CI jobs to Nodepool - https://phabricator.wikimedia.org/T119143#2170445 (10greg) [20:56:58] !log Refreshing Nodepool snapshot to hopefully get npm 2.x installed T124474 [20:56:58] T124474: Package npm 2.14 - https://phabricator.wikimedia.org/T124474 [20:57:03] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [20:57:18] hashar: :) [20:57:48] (03PS3) 10Hashar: Update npm to 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280912 (owner: 10Paladox) [20:59:12] (03Abandoned) 10Hashar: Update npm to 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280912 (owner: 10Paladox) [20:59:29] (03CR) 10Hashar: [C: 032] dib: provision npm 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280980 (https://phabricator.wikimedia.org/T124474) (owner: 10Hashar) [20:59:30] PROBLEM - Puppet run on deployment-tin is CRITICAL: CRITICAL: 20.00% of data above the critical threshold [0.0] [20:59:33] which me luck [20:59:40] whish [20:59:42] whatever [20:59:43] 10Continuous-Integration-Config, 5Continuous-Integration-Scaling, 10releng-201516-q3, 3releng-201516-q4, 7WorkType-NewFunctionality: [keyresult] Migrate php composer (Zend and HHVM) CI jobs to Nodepool - https://phabricator.wikimedia.org/T119139#2170466 (10greg) Added our q4 project to this [20:59:50] paladox: so yeah npm install npm@2.x would do [20:59:53] UIDGeneratorTest::testTimestampedUID with data set #2 ('newTimestampedUID88', 27, 88, 46, 32) [20:59:53] 20:49:19 New ID timestamp (10101001111010011100101100011111011011110000000000000100011100010100010111001110101) >= prior one (10101001111010011100101100011111011011110111111111100100011100010100010111001110101). [20:59:53] 20:49:19 Failed asserting that '1010100111101001110010110001111101101111000000' is equal to or is greater than '1010100111101001110010110001111101101111011111'. [20:59:56] * OuKB wishes hashar whatever [20:59:57] wtf is up with jenkins? [21:00:05] paladox: but it is far easier to have the node pool instance to already come with the proper npm version :-} [21:00:07] hashar: cool. So how does the image update work actually? [21:00:15] hashar: Yep [21:00:20] Reedy: jenkins just run whatever dev defined. So really it is most probably an issue within the code [21:00:28] How do you push from integration/config to the base image changing [21:00:33] Reedy: or the lame MediaWiki timestamp generator being crazy [21:00:33] (03Merged) 10jenkins-bot: dib: provision npm 2.x [integration/config] - 10https://gerrit.wikimedia.org/r/280980 (https://phabricator.wikimedia.org/T124474) (owner: 10Hashar) [21:01:23] Krinkle: Nodepool uses a specially crafted images in OpenStack labs. Once per day around 14:00 UTC, it would boot an instance out of that image, git pull puppet and integration/config then run the puppet script in integration/config.git /dib [21:01:29] Krinkle: that is all automatic [21:01:45] cool [21:01:52] twentyafterfour: thcipriani, woohoo! it works! needs a little more sculpting, but it works in beta! [21:01:54] Krinkle: and we can ask nodepool to do the process [21:01:59] https://wikitech.wikimedia.org/wiki/Nodepool#Manually_generate_a_new_snapshot [21:02:09] become-nodepool && git -C /etc/nodepool/wikimedia/ pull && nodepool image-update wmflabs-eqiad ci-jessie-wikimedia [21:02:22] want me to screen cast my terminal ? :-} [21:02:32] hashar: You know export we did for bundle. Would we be able to do it for npm 0.10 please otherwise one using npm 0.10 and the other nodejs 4.3 will break. [21:02:38] ottomata: nice :) [21:03:06] Krinkle: paladox: nodepool update at https://plus.google.com/hangouts/_/wikimedia.org/nodepoolupdate?authuser=0 if you wanna lurk :-} [21:04:07] paladox: yeah gotta think about it :} [21:04:24] hashar: Ok [21:08:22] hashar: Are we updating nodejs to 4.4 at the same time as updating npm to 2.x [21:09:40] 10Continuous-Integration-Config, 5Continuous-Integration-Scaling, 10releng-201516-q3, 3releng-201516-q4, 7WorkType-NewFunctionality: [keyresult] Migrate php composer (Zend and HHVM) CI jobs to Nodepool - https://phabricator.wikimedia.org/T119139#2170490 (10Paladox) Maybe we can update php using the nodep... [21:09:45] RECOVERY - Host deployment-upload is UP: PING OK - Packet loss = 0%, RTA = 0.78 ms [21:09:55] paladox: Unlikely. Node upgrades are more impactful in production, that'll be done once a year together with Debian and Ubuntu upgrades. [21:10:10] Krinkle ok, thanks for explaning [21:10:23] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2170491 (10Andrew) [21:13:45] !log Image ci-jessie-wikimedia-1459544873 in wmflabs-eqiad is ready [21:13:51] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [21:14:06] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2170499 (10Krenair) 5Open>3Resolved a:3Andrew [21:14:42] PROBLEM - Puppet staleness on deployment-upload is CRITICAL: CRITICAL: 33.33% of data above the critical threshold [43200.0] [21:18:02] Krinkle or hashar do you know why it keeps failing with errors like https://integration.wikimedia.org/ci/job/mediawiki-extensions-qunit/37330/console i have been seeing alot of these errors lately. [21:19:12] 10Beta-Cluster-Infrastructure: /mnt/upload7 does not exist anywhere, yet it is referenced in multiple places in wmf-config - https://phabricator.wikimedia.org/T129586#2170532 (10greg) [21:20:42] ottomata: awesome [21:22:56] 10Beta-Cluster-Infrastructure, 6Commons, 10MediaWiki-File-management, 6Multimedia: Thumbnail generation should happen via the same setup in the beta cluster and in production (tracking) - https://phabricator.wikimedia.org/T84950#2170538 (10greg) >>! In T84950#2168051, @Tgr wrote: > Video also seems broken... [21:24:04] 10Beta-Cluster-Infrastructure, 6Labs: deployment-upload won't start, upload.beta.wmflabs.org down - https://phabricator.wikimedia.org/T131322#2163472 (10greg) >>! In T131322#2170491, @Andrew wrote: > YOUR job is to puppetize the damn thing :) Just for the record, that's mostly blocked on (afaiui): * {T64835}... [21:24:17] hashar: Hi could you remove npm verb from https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/10/console please. [21:24:26] 5Continuous-Integration-Scaling, 3releng-201516-q4, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate javascript npm CI jobs to Nodepool - https://phabricator.wikimedia.org/T119143#2170546 (10greg) [21:24:44] RECOVERY - Puppet staleness on deployment-upload is OK: OK: Less than 1.00% above the threshold [3600.0] [21:28:09] Krinkle or hashar is it posible on trusty to update nodejs to 0.10.44 since nodejs 0.10.25 seems buggy. Plus outdated [21:28:19] please [21:28:31] hashar: It passes https://integration.wikimedia.org/ci/job/parsoidsvc-source-npm-node-4.3/10/console [21:34:24] 5Gerrit-Migration, 10Differential, 7Documentation: Document use of Owners in Phabricator and advertise it - https://phabricator.wikimedia.org/T128372#2072003 (10greg) One example use of the Owners application is catching commits that weren't reviewed in Differential pre-merge (aka "cowboy commits") thus bypa... [21:36:07] PROBLEM - Puppet run on integration-slave-trusty-1018 is CRITICAL: CRITICAL: 22.22% of data above the critical threshold [0.0] [21:40:20] paladox: would need a debian package to be build [21:40:40] paladox: and I dont think Ubuntu or anyone will provide an update of node package for Trusty (but I may be wrong) [21:40:59] Ok [21:41:12] http://packages.ubuntu.com/trusty/nodejs it is in 'universe' [21:41:15] so community maintained [21:41:27] i.e. it is most definitely never going to be updated [21:43:58] 5Continuous-Integration-Scaling, 10Parsoid, 6Services, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate Parsoid CI jobs from node 0.8/0.10 to 4.3 - https://phabricator.wikimedia.org/T126992#2170617 (10hashar) [21:44:02] 10Continuous-Integration-Infrastructure, 6Operations, 6Services, 13Patch-For-Review: Package npm 2.14 - https://phabricator.wikimedia.org/T124474#2170614 (10hashar) 5stalled>3Resolved a:3hashar Nodepool images now comes with npm 2.5.12. provisioned by puppet and installed from npmjs.org. ``` $ /usr/... [21:44:21] 5Continuous-Integration-Scaling, 6Services, 7Tracking, 7WorkType-NewFunctionality: [tracking] Migrate services CI jobs to node 4.3 - https://phabricator.wikimedia.org/T126991#2170619 (10hashar) [21:44:26] 5Continuous-Integration-Scaling, 10Parsoid, 6Services, 13Patch-For-Review, 7WorkType-NewFunctionality: Migrate Parsoid CI jobs from node 0.8/0.10 to 4.3 - https://phabricator.wikimedia.org/T126992#2029093 (10hashar) 5stalled>3Open [21:44:41] paladox: nodepool has npm 2.x now :) [21:44:58] hashar: Thanks :) [21:45:04] paladox: "Buggy" in what way? What bugs did you encounter? [21:45:45] gotta sleep now. Will get to Parsoid next week I guess :-} [21:45:50] have a good week-end [21:45:53] Krinkle well it seems most packages need nodejs 0.10 and it seems we keep getting package.json cannot be opened. [21:45:57] ok [21:46:20] hashar: What should we do to export the bundle path on npm 0.10. [21:47:07] paladox: evaluate what side effect it is going to have to other jobs using bundle install :-} [21:47:13] such as the browser test jobs [21:47:18] hashar: Oh ok [21:48:48] paladox: that will be for next week :-} [21:48:56] ok [21:51:06] paladox: one sure thing, thank you very much for all the patches this week. Special mention to have get rid of the mwext--testextension-* jobs :-D [21:51:18] sleep weel [21:51:23] well [21:51:31] :P [21:51:53] yeah [21:51:57] aheo [21:52:00] g'night :) [21:52:22] I wich I cloud wright anglish properly [21:52:55] hashar: You should have seen old english and middle age english. Nothing like it is today. [21:58:47] paladox: same deal with french :D [21:58:58] I disappear for real [22:04:17] (03CR) 10Paladox: "Thanks." [integration/config] - 10https://gerrit.wikimedia.org/r/280940 (owner: 10Paladox) [22:08:22] 6Release-Engineering-Team: MW-1.27.0-wmf.21 deployment blockers - https://phabricator.wikimedia.org/T131555#2170670 (10greg) [22:08:36] 6Release-Engineering-Team: MW-1.27.0-wmf.20 deployment blockers - https://phabricator.wikimedia.org/T130945#2151605 (10greg) [22:09:38] 10Continuous-Integration-Config: Whitelist Lokal Profil for CI - https://phabricator.wikimedia.org/T131558#2170722 (10JeanFred) [22:10:05] 6Release-Engineering-Team: MW-1.27.0-wmf.24 deployment blockers - https://phabricator.wikimedia.org/T131559#2170735 (10greg) [22:10:20] 6Release-Engineering-Team: MW-1.27.0-wmf.23 deployment blockers - https://phabricator.wikimedia.org/T131557#2170705 (10greg) [22:10:28] 6Release-Engineering-Team: MW-1.27.0-wmf.22 deployment blockers - https://phabricator.wikimedia.org/T131556#2170691 (10greg) [22:12:44] paladox: Node 0.10 can open a package.json just fine, and has for over 3 years. [22:12:49] Thebug is not with Node.js [22:13:01] Krinkle oh ok [22:13:45] If Node 0.10 was so broken that it could not open a simple json file, then it would have never passed unit test in nodejs/node and would not have been released. [22:13:48] (03PS1) 10Jean-Frédéric: Add Lokal Profil to trusted users [integration/config] - 10https://gerrit.wikimedia.org/r/281046 (https://phabricator.wikimedia.org/T131558) [22:14:33] (03CR) 10Krinkle: [C: 032] "I'm not biased." [integration/config] - 10https://gerrit.wikimedia.org/r/281046 (https://phabricator.wikimedia.org/T131558) (owner: 10Jean-Frédéric) [22:14:48] 10Continuous-Integration-Config, 13Patch-For-Review: Whitelist Lokal Profil for CI - https://phabricator.wikimedia.org/T131558#2170763 (10JeanFred) [22:16:05] (03Merged) 10jenkins-bot: Add Lokal Profil to trusted users [integration/config] - 10https://gerrit.wikimedia.org/r/281046 (https://phabricator.wikimedia.org/T131558) (owner: 10Jean-Frédéric) [22:16:07] RECOVERY - Puppet run on integration-slave-trusty-1018 is OK: OK: Less than 1.00% above the threshold [0.0] [22:16:46] !log Reloading Zuul to deploy https://gerrit.wikimedia.org/r/281046 [22:16:51] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL, Master [22:17:04] 10Continuous-Integration-Config, 13Patch-For-Review: Whitelist Lokal Profil for CI - https://phabricator.wikimedia.org/T131558#2170769 (10Krinkle) 5Open>3Resolved p:5Triage>3Normal [22:17:22] 10Continuous-Integration-Config, 13Patch-For-Review: Whitelist Lokal Profil for CI - https://phabricator.wikimedia.org/T131558#2170722 (10Krinkle) @Lokal_Profil Welcome! [22:57:53] 6Release-Engineering-Team, 6Developer-Relations, 6Team-Practices, 15User-greg: Set up Code Review office hours - https://phabricator.wikimedia.org/T128371#2170876 (10greg) If anyone watching this task is at the Hackathon right now: Feel free to talk to more people there about it and get them to vote/commen... [23:05:43] (03PS3) 10Paladox: Export BUNDLE_PATH to npm and npm-0.10 tests [integration/config] - 10https://gerrit.wikimedia.org/r/280929 [23:07:04] 6Release-Engineering-Team: MW-1.27.0-wmf.19 deployment blockers - https://phabricator.wikimedia.org/T130227#2170888 (10greg) 5Open>3Resolved [23:12:25] paladox: twentyafterfour: ping? [23:12:39] Hi [23:14:14] twentyafterfour: Hi, Dereckson help me to get arc to work on the server he let me used but me and him are having problems uploading. We are getting to the bit where it says pushing staging. But then it asks for our username and password when it should just use our ssh pub key. Plus pushing to staging is failing even adding the ssh key is not working. [23:16:10] twentyafterfour: so, currently staging configuration for rMSCA uses https, not SSH [23:16:52] that would be more convenient if it uses SSH, to have a seamless auth without any need of password [23:27:04] 10Deployment-Systems, 6Release-Engineering-Team: Where should we branch for Wikimedia wikis? - https://phabricator.wikimedia.org/T121570#2170960 (10greg) @demon / @thcipriani ya'll have been doing this on a labs instance lately, right? (Branching mw+extensions pre-train) [23:27:32] greg-g: or twentyafterfour: COuld you setup a repo on phabricator.wikimedia.org or the one that we use for testing so we can test arc land please. [23:27:45] Dereckson ^^ [23:29:04] 10Deployment-Systems, 6Release-Engineering-Team, 13Patch-For-Review, 15User-greg, 7user-notice: Move the train deployment from Thursday to Wednesday for some Wikipedia sites - https://phabricator.wikimedia.org/T115002#2170962 (10greg) 5Open>3Resolved a:3greg >>! In T115002#1877919, @Amire80 wrote:... [23:29:18] paladox: I'm not sure why staging doesn't work? you set up ssh keys in phabricator? [23:29:29] twentyafterfour: Works now [23:29:40] But it dosent want your ssh key it wants a vcs password [23:29:58] paladox: let's use https://phab-01.wmflabs.org/diffusion/ for that as well [23:30:08] twentyafterfour: I can clone ssh://vcs@git-ssh.wikimedia.org, but at staging operation it offers me http too. [23:30:23] we don't want to use production phabricator for test repos/tasks/diffs [23:30:25] twentyafterfour: How can we get it so if we use ssh staging will use ssh [23:30:43] greg-g: Ok [23:30:55] greg-g: there isn't any differential on wmflabs, afaik [23:30:58] twentyafterfour: staging is maybe set in https only at https://phabricator.wikimedia.org/diffusion/MSCA/edit/ ? [23:31:09] Dereckson: no it's set to ssh or https [23:31:46] Oh, but its enforcing http for staging [23:32:35] Dereckson: paladox: Figured it out. it was the scap repo that had the staging url set to https [23:32:37] fixed it [23:32:38] twentyafterfour: we speak about the Staging Area URI [23:32:45] and only one url is possible [23:32:54] okay, cool [23:32:54] twentyafterfour: Thanks [23:33:04] Could you check the repos considered for Phase I? [23:33:19] Dereckson: what do you mean? [23:33:31] These ones: https://phabricator.wikimedia.org/T130418 [23:33:47] most repos don't have a staging setting yet [23:33:54] and we may not end up using the staging repo at all [23:34:02] ok [23:34:10] I'll take a lok [23:34:12] look [23:34:37] twentyafterfour: hrmmm, how hard is it to setup there? mayve a test repo in production phab is better? [23:35:24] Works now [23:35:35] greg-g: not too easy to set up [23:35:49] 10Deployment-Systems, 6Release-Engineering-Team: Where should we branch for Wikimedia wikis? - https://phabricator.wikimedia.org/T121570#2170974 (10thcipriani) >>! In T121570#2170960, @greg wrote: > @demon / @thcipriani ya'll have been doing this on a labs instance lately, right? (Branching mw+extensions pre-t... [23:36:31] twentyafterfour and greg-g is there any way we could setup a rep that we can do arc land without causing any harm to any one. [23:36:57] paladox: yes [23:37:07] * twentyafterfour sets one up [23:37:10] twentyafterfour: Thanks :) [23:38:12] 10Deployment-Systems, 6Release-Engineering-Team: Where should we branch for Wikimedia wikis? - https://phabricator.wikimedia.org/T121570#2170975 (10greg) 5Open>3Resolved a:3thcipriani the deploy host it is [23:40:17] paladox: https://phabricator.wikimedia.org/diffusion/1883/ [23:40:31] twentyafterfour: Thanks [23:41:07] paladox: can you abandon the other test diffs in the scap repo when you're done with them? Thanks! [23:41:18] Ok [23:45:09] twentyafterfour: Could you add the arcconfig file to https://phabricator.wikimedia.org/diffusion/1883/ please so that when we are doing arc install-certificate it will go to the correct link. [23:47:38] 10Deployment-Systems, 6Project-Admins, 15User-greg: Further cleanup of #Deployment-Systems - https://phabricator.wikimedia.org/T126631#2019200 (10greg) [23:48:22] 10Deployment-Systems, 6Project-Admins, 15User-greg: Further cleanup of #Deployment-Systems - https://phabricator.wikimedia.org/T126631#2019200 (10greg) //(I updated the description with the current working theory as I had in my head.)// [23:56:58] twentyafterfour and greg-g could we create a task for create the .arcconfig file for all repos. Possibly we can do it now so for new gerrit repos being created and phabricator. There should be .arcconfig created per Dereckson. [23:57:05] please [23:58:21] https://phabricator.wikimedia.org/T130787 [23:58:22] :) [23:58:30] greg-g: Thanks. [23:58:39] I can add a note to the gerrit creations page [23:59:49] 10Continuous-Integration-Infrastructure, 5Continuous-Integration-Scaling, 10releng-201415-Q3, 10releng-201415-Q4, and 2 others: [EPIC] Run CI jobs in disposable VMs - https://phabricator.wikimedia.org/T47499#2171035 (10Krinkle) [23:59:51] 10Continuous-Integration-Infrastructure: Consolidate jobs to test entry points - https://phabricator.wikimedia.org/T91996#2171033 (10Krinkle) 5Open>3Resolved a:3Krinkle