[01:27:50] @greg-g: Hey greg [02:56:10] kaldari: ? [07:36:48] 10Phabricator, 10Release-Engineering-Team (Kanban): searching "db1055" on phabricator search doesn't find the only open task with "db1055" on the title - https://phabricator.wikimedia.org/T194154#4193215 (10jcrespo) Finding the mentioned task now works for me. [07:37:45] 10Phabricator, 10Wikimedia-Site-requests, 10Patch-For-Review, 10Wikimedia-Incident: Cleanup phabricator.wikimedia.org uploaded files, WP zero abuse - https://phabricator.wikimedia.org/T168142#3361850 (10Addshore) I ran into this block yesterday while working from Morocco, I'm not sure if the block might be... [07:50:55] 10Phabricator, 10Wikimedia-Site-requests, 10Patch-For-Review, 10Wikimedia-Incident: Cleanup phabricator.wikimedia.org uploaded files, WP zero abuse - https://phabricator.wikimedia.org/T168142#4193239 (10Aklapper) For a bigger picture, see also: {T173537} and {T185784} . Zero data might be wrong sometimes. [08:03:38] 10Phabricator, 10Wikimedia-Site-requests, 10Patch-For-Review, 10Wikimedia-Incident: Cleanup phabricator.wikimedia.org uploaded files, WP zero abuse - https://phabricator.wikimedia.org/T168142#4193257 (10Paladox) We will be able to remove the Wikipedia zero block this year as Wikipedia zero is being discont... [08:46:05] 10Continuous-Integration-Config, 10MediaWiki-extensions-AdvancedMeta: Re-open repo AdvancedMeta - https://phabricator.wikimedia.org/T194242#4193299 (10MarcoAurelio) Actually I think it's best for @hashar to do so given that he's working on migrating extensions to a new checking system and I don't know which of... [08:46:45] 10Continuous-Integration-Config, 10MediaWiki-extensions-AdvancedMeta: Restore jenkins-bot normal tests (remove "archived") for extension AdvancedMeta - https://phabricator.wikimedia.org/T194242#4193302 (10MarcoAurelio) p:05Triage>03Normal [09:32:43] 10Release-Engineering-Team (Kanban), 10Scap, 10Scoring-platform-team, 10Patch-For-Review: Support git-lfs - https://phabricator.wikimedia.org/T180627#4193364 (10demon) >>! In T180627#4192775, @awight wrote: > 15:50 < awight> twentyafterfour: bad news, my test LFS deployment failed to pull the files again >... [09:32:45] 10Gerrit, 10Developer-Relations, 10artificial-intelligence: Patch-wrangler -- suggests the best reviewers for a patch - https://phabricator.wikimedia.org/T155851#4193365 (10Aklapper) [09:37:18] (03PS1) 10MarcoAurelio: Whitelist Rxy on zuul [integration/config] - 10https://gerrit.wikimedia.org/r/432051 [09:41:10] (03PS2) 10MarcoAurelio: Add AbuseFilter dependency on TorBlock [integration/config] - 10https://gerrit.wikimedia.org/r/431806 [10:15:28] no_justification: there’s the gerrit project in cloud we could use to test 2.14 -> 2.15 (without touching gerrit-test3) :) [10:15:37] So basically two instances that are like prod [10:16:22] One will be on the same update schedule as prod whilst we test 2.15 on another instance that’s like prod [10:16:27] Testing puppet changes [13:01:14] 10Phabricator, 10Release-Engineering-Team (Kanban): searching "db1055" on phabricator search doesn't find the only open task with "db1055" on the title - https://phabricator.wikimedia.org/T194154#4193905 (10mmodell) 05Open>03Resolved [13:24:33] 10Release-Engineering-Team, 10Wikimedia-Hackathon-2018: Work on "Docker environment generator" - https://phabricator.wikimedia.org/T192789#4193968 (10Mholloway) Out of curiosity, how does all of this relate to the #mediawiki-containers project (https://github.com/wikimedia/mediawiki-containers), if at all? [13:27:49] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Different production servers have different versions of tidy installed, resulting in varying output - https://phabricator.wikimedia.org/T193414#4193987 (10ssastry) 05Open>03declined [13:28:25] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MW-1.32-release: Expand the set of bundled skins - https://phabricator.wikimedia.org/T194266#4193988 (10CCicalese_WMF) [13:30:18] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MW-1.31-release, 10Patch-For-Review: Expand the set of bundled extensions to achieve a default MediaWiki experience that's comparable to Wikimedia sites - https://phabricator.wikimedia.org/T178349#4194002 (10CCicalese_WMF) I created T194266 to continue the di... [13:46:21] 10Beta-Cluster-Infrastructure: Request for shell access and steward rights in beta cluster - https://phabricator.wikimedia.org/T194267#4194035 (10Rxy) [13:52:44] 10Continuous-Integration-Config, 10MediaWiki-extensions-AdvancedMeta: Restore jenkins-bot normal tests (remove "archived") for extension AdvancedMeta - https://phabricator.wikimedia.org/T194242#4194049 (10hashar) Just revert the change that archived it 14642fd72f92213e3bb9c5632c2c3058d8d07022 or namely add: ``... [14:27:46] thcipriani: For some security updates I'm trying to choose a labvirt host to use as a lab rat… my leading candidate hosts 'deployment-etcd-01' and 'deployment-ms-be03'. How much of a mess will it cause you if those reboot a few times unexpectedly? [14:28:00] Well, or (worst-case scenario) stop working for hours in the middle of the night [14:34:15] oh boy. My guess would be etcd-01 is used by mediawiki to get config so beta would probably stop serving. deployment-ms-be03 is part of the beta swift cluster and I'm unsure of what would happen if that one went down, but I suppose I'd lean towards that one. [14:34:44] since not every request would try to connect to it [14:37:08] thcipriani: sorry, my grammar was unclear — the host I want to reboot has both those VMs on it. [14:37:17] oh :) [14:37:26] I could migrate them elsewhere before I start my testing, but that would probably produce longer downtime than just rebooting [14:37:30] well in that case beta will be down :P [14:37:34] ok [14:37:44] hm [14:38:16] Is there any time of day that's better or worse for that to happen? (Not going to get to this today in any case, can schedule ahead) [14:40:33] really the hour or two around now seems to usually be fairly quiet. Or, roughly, 5pm-10pm pacific. These are the times that I've seen problems go unnoticed for long periods anyway. [14:41:25] Hm… maybe I /should/ do it right now then... [14:41:34] that would give it all day to fail unexpectedly before I go to sleep [14:41:48] \o/ [14:42:02] thcipriani: what do you think? Should I just start breaking things now? Or would you rather by appointment in a few days? [14:43:17] Anyone an idea why https://gerrit.wikimedia.org/r/#/c/432049/ is not working? [14:43:19] I tried to use the vfsstramer lib installed with composer in a phpunit test [14:43:53] if it's all the same to you, I'd rather in a few days, I'd like to attempt a jenkins update Soon™ and I'd like that to happen with some isolation. If maybe just sanity isolation :) [14:44:20] thcipriani: no problem — I'll schedule something and send you a calendar invite [14:44:28] cool, thanks [14:45:19] is Friday too soon? Or Monday? [14:45:54] Monday releng is starting our offsite, Friday would work [14:46:10] great [14:48:50] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MW-1.32-release: Expand the set of bundled skins - https://phabricator.wikimedia.org/T194266#4193988 (10RazeSoldier) MW 1.31 already bundled with Timeless? [14:50:37] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MW-1.32-release: Expand the set of bundled skins - https://phabricator.wikimedia.org/T194266#4194268 (10Legoktm) [14:50:58] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MW-1.32-release: Expand the set of bundled skins - https://phabricator.wikimedia.org/T194266#4193988 (10Legoktm) >>! In T194266#4194265, @RazeSoldier wrote: > MW 1.31 already bundled with Timeless? Yes, I updated the description to be more explicit about that. [15:27:57] 10Beta-Cluster-Infrastructure, 10Release-Engineering-Team (Next), 10Patch-For-Review: Beta puppetmaster cherry-pick process - https://phabricator.wikimedia.org/T135427#4194376 (10Krinkle) > * 023dd6c75a (5 months ago) Timo Tijhof - varnish: Add --errorpage-noise trigger for easy testing I've removed this on... [15:39:34] Project beta-scap-eqiad build #207060: 04STILL FAILING in 3 min 25 sec: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/207060/ [16:04:02] > error writing ‘/tmp/scap_l10n_3038535146/l10n_cache-sr.cdb’: No space left on device [16:06:57] I cleared some space, waiting to see if it works then I'll log what I done [16:07:24] sorry, should have claimed it [16:07:32] :) [16:13:01] 10Phabricator: Create Tag for WMDE Fundraising Security issues - https://phabricator.wikimedia.org/T194286#4194474 (10gabriel-wmde) [16:17:04] !log deployment-tin:apt-get clean && rm -rf /srv/mediawiki/.git to free space [16:17:06] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [16:17:16] (looks like I'm going to be waiting a while :( ) [16:18:44] (03CR) 10Umherirrender: [C: 031] Add AbuseFilter dependency on TorBlock [integration/config] - 10https://gerrit.wikimedia.org/r/431806 (owner: 10MarcoAurelio) [16:26:29] Yippee, build fixed! [16:26:30] Project beta-scap-eqiad build #207061: 09FIXED in 42 min: https://integration.wikimedia.org/ci/job/beta-scap-eqiad/207061/ [16:34:26] yay [17:15:39] 10Phabricator, 10Project-Admins: Create Tag for WMDE Fundraising Security issues - https://phabricator.wikimedia.org/T194286#4194620 (10Framawiki) [17:24:06] O/ greg-g [17:24:12] ahoy hoy [17:24:25] So, I loved the whole wikibase lexeme timeline forward 1 week btw :) (since I last pinged you) fyi [17:24:39] s/loved/moved/ ? [17:24:53] Heh, moved, stupid phone [17:25:07] a lovely phone [17:25:12] so, next week to test.wikidata? [17:25:19] Yup [17:25:23] And the plan is beta tonight for me [17:25:32] I'll update the wiki since you're on your phone [17:25:40] I already updated the wiki :) [17:25:47] But not scheduled any exact slots yet [17:25:55] oh, you did [17:26:04] :) [17:26:39] It's fine for me to schedule slots for me to deploy stuffbisbgone EU morning and datytime I guess? :) There is no limit to where I am allowed to schedule then? [17:27:27] stuffbisbgone? what is your phone on about? [17:27:55] Haha, *deploy stuff in EU mounting... [17:27:58] Morning, ffs [17:29:03] 10Release-Engineering-Team (Kanban), 10Scap, 10Scoring-platform-team, 10Patch-For-Review: Support git-lfs - https://phabricator.wikimedia.org/T180627#4194644 (10awight) > You know, we should probably install git-lfs everywhere we can, just like git, and do `git lfs install --global` as part of it. That wo... [17:31:51] addshore: yeah, you'll be fine [17:32:05] we'll be around, but not too much on our laptops, next week, during EU mornings [17:40:47] 10Release-Engineering-Team (Kanban), 10Scap, 10Operations, 10Patch-For-Review: mwscript rebuildLocalisationCache.php takes 40 minutes - https://phabricator.wikimedia.org/T191921#4121650 (10Paladox) coulden't a new flag be added that switches between php7.0 and hhvm? [17:54:59] greg-g: hey, there's no deployment next week due to your offsite, right? [17:55:36] looking at https://www.mediawiki.org/wiki/MediaWiki_1.32/Roadmap [17:55:43] no_justification: fyi, my git mirroring failed in a new way, https://phabricator.wikimedia.org/source/drafttopic/manage/uris/ [17:55:55] The phabricator import completed, but nothing was pushed to gerrit. [17:56:27] Might it help if I merge a dummy commit in github now, i.e. maybe the mirroring job failed and I just need to kick off another? [18:02:48] Volker_E: don't trust that, the canonical source for deployment schedule is https://wikitech.wikimedia.org/wiki/Deployments. Next week the train is on (mukunda and chad are not going to the offsite) [18:24:24] twentyafterfour: The services deployment window overlaps with the MW train today, that seems like an accident and we shouldn't do services things, correct? [18:29:20] 10Release-Engineering-Team, 10Scoring-platform-team (Current): Phabricator repo is not mirroring to gerrit as hoped - https://phabricator.wikimedia.org/T194295#4194783 (10awight) [18:29:40] 10Release-Engineering-Team, 10Scoring-platform-team (Current): Phabricator repo is not mirroring to gerrit as hoped - https://phabricator.wikimedia.org/T194295#4194793 (10awight) [18:32:51] greg-g: oh cool, where will you be? offsite pre hackathon? [18:32:55] / post hackathon [18:33:20] awight: let me take a look, I'm used with Diffusion setup [18:33:21] addshore: pre, in barcelona itself. [18:33:28] sweet :) [18:33:46] I'll be seeing you there for the hackathon, unfortunatly I havn't written enough browser tests to deserve any kind of reward though [18:33:59] Hauskatze: thank you! [18:34:09] * addshore is currently doing his own single man offsite in Morocco [18:34:53] awight: so where is the repo getting the code from, github? [18:35:02] ye [18:35:03] yes [18:35:14] https://phabricator.wikimedia.org/source/drafttopic/manage/uris/ [18:35:33] I've been able to confirm that the observing from GitHub->Phabricator works correctly [18:36:03] so it's the diffusion --> gerrit the one not working? [18:36:51] exactly [18:37:16] hmm, maybe the go-import plugin not correctly setup no_justification ? [18:37:46] ....go-import? [18:37:58] Doesn't have anything to do with the mirroring..... [18:38:04] https://gerrit.wikimedia.org/r/#/admin/plugins/ [18:38:16] it's because you have it set to anon https url [18:38:16] https://phabricator.wikimedia.org/source/drafttopic/uri/view/19929/ [18:38:35] you need to use the account that's setup to push to gerrit. [18:38:46] well the uri is wrong https://gerrit.wikimedia.org/r/scoring/ores/drafttopic [18:38:47] It needs K18, right? [18:38:50] 404 [18:39:03] yep [18:39:05] seems so [18:39:09] yep [18:39:27] So fix the URL if needed, apply K18, then it'll work [18:39:51] k18 applied [18:40:17] and url fixed as well afaics [18:41:24] https://gerrit.wikimedia.org/g/scoring/ores/drafttopic still empty, I guess it'll take a while [18:43:10] no_justification does the user have access to all repos on gerrit? [18:43:12] ie phab? [18:43:17] (the user named phab) [18:43:20] can it push to them [18:45:21] paladox: I can't see any user named as such but a mmodell account named "Phabricator" with a @wikimedia address [18:45:35] Hauskatze on gerrit? [18:45:41] yep [18:45:49] Hauskatze the user is named phab [18:45:52] mmodell+phab [18:46:03] looking for owner:phab is the only result I get [18:47:26] 10Gerrit: mediawiki/extensions repo lacks extensions - https://phabricator.wikimedia.org/T194156#4190450 (10Umherirrender) Should be rewritten to use the rest api https://gerrit.wikimedia.org/r/projects/?d&format=json&p=mediawiki%2Fextensions%2F ``` extensions PreferencesMaster missing extensions ShowRealUser... [18:47:39] hmm [18:47:41] https://gerrit.wikimedia.org/r/#/q/owner:%22Phabricator+%253Cmmodell%252Bphab%2540wikimedia.org%253E%22 [18:47:52] yes, that's the one [18:48:40] hmm [18:49:30] Hauskatze yep that's the account [18:49:31] https://gerrit.wikimedia.org/r/#/admin/groups/1217,members [18:52:37] Hauskatze: for comparison, this one seems to work https://phabricator.wikimedia.org/source/editquality/manage/uris/ [18:52:52] (03PS1) 10MarcoAurelio: Modify access rules [scoring/ores] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432141 [18:52:52] & it has the same style gerrit URL as drafttopic had... [18:53:18] awight: can you approve the above and see if that works? [18:53:54] awight: wait [18:53:56] wrong repo [18:54:00] (03Abandoned) 10MarcoAurelio: Modify access rules [scoring/ores] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432141 (owner: 10MarcoAurelio) [18:54:44] aha! [18:57:26] (03PS1) 10MarcoAurelio: Modify access rules [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432143 [18:57:37] awight: ^^ [18:57:44] kk [18:58:58] (03CR) 10Awight: [C: 032] "Great find, thank you!" [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432143 (owner: 10MarcoAurelio) [18:59:08] (03CR) 10Paladox: [V: 032] Modify access rules [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432143 (owner: 10MarcoAurelio) [18:59:16] apparently i can v+2 lol [18:59:23] verily [18:59:26] anyways awight you need to press submit on that [18:59:40] yep [18:59:51] I could have c2v2 anyway, but not submit [18:59:57] Done. That's a huge help! [18:59:58] i think we should just grant that in All-Projects [19:00:02] push rights for phab [19:00:04] no_justification ^^ [19:00:09] awight: now lets see if that works [19:00:59] I suppose I should merge something in GH to kick it off? [19:01:32] idk, but I'd try that I guess [19:02:04] 10Release-Engineering-Team (Kanban), 10Scap, 10Operations, 10Patch-For-Review: mwscript rebuildLocalisationCache.php takes 40 minutes - https://phabricator.wikimedia.org/T191921#4194970 (10Legoktm) >>! In T191921#4161858, @Krinkle wrote: > I've put a straw-man up at T176370#4161855. Seems reasonable, but... [19:02:53] (03PS1) 10Paladox: Modify access rules [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 [19:03:35] (03PS2) 10Paladox: Modify access rules [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 [19:03:38] awight Hauskatze ^^ [19:03:49] that's what editquality does [19:04:14] that's needed anyways [19:04:18] needs the forge right [19:04:19] too [19:04:23] for the git mirror thing [19:04:24] (03CR) 10Awight: [V: 032 C: 032] "Cool, let's try this." [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 (owner: 10Paladox) [19:04:32] thanks awight :) [19:05:06] ah :) [19:05:36] (03CR) 10MarcoAurelio: "You need to submit :)" [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 (owner: 10Paladox) [19:05:45] awight ^^ [19:05:52] need to press the submit button :) [19:05:57] refs/meta/* dosen't have ci [19:06:33] Apparently I can't submit, we need a gerrit admin [19:06:40] oh [19:06:50] yep, changing inherits requires a phab admin [19:07:23] no_justification ^^ [19:07:48] though we could just replicate it without changing parent project? [19:07:51] awight thoughts ^^? [19:08:38] Seems like doing it the kosher way is worth a small delay [19:09:02] ok [19:09:08] does changing inheritance means people loosing access to the repo? [19:09:15] because that repo has no owner [19:09:36] Hauskatze i set research-ores as the owner, so i am guessing that it's merging owners [19:09:45] kk [19:09:46] oh [19:09:47] wait [19:09:50] submit right [19:10:30] (03PS1) 10Paladox: test [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432149 [19:10:55] (03PS3) 10Paladox: Modify access rules [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 [19:10:59] (03Abandoned) 10Paladox: test [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432149 (owner: 10Paladox) [19:11:06] awight https://gerrit.wikimedia.org/r/#/c/432145/ [19:11:41] (03CR) 10Awight: [V: 032 C: 032] Modify access rules [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 (owner: 10Paladox) [19:12:38] (03CR) 10Awight: [V: 032 C: 032] "I'm not sure we want submit on this repo, since it's a mirror..." [scoring/ores/drafttopic] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/432145 (owner: 10Paladox) [19:13:23] (03PS1) 10MarcoAurelio: Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 [19:13:34] (03PS2) 10MarcoAurelio: Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 [19:15:18] (03PS3) 10MarcoAurelio: Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 (https://phabricator.wikimedia.org/T194242) [19:15:51] 10Phabricator, 10Project-Admins, 10Security-Team: Create Tag for WMDE Fundraising Security issues - https://phabricator.wikimedia.org/T194286#4194998 (10Aklapper) Simplified speaking: There are Security tickets (non-public; can be accessed by everyone who is in https://phabricator.wikimedia.org/project/membe... [19:16:08] 10Gerrit, 10Release-Engineering-Team, 10Scoring-platform-team (Current): Phabricator repo is not mirroring to gerrit as hoped - https://phabricator.wikimedia.org/T194295#4195003 (10awight) We need a Gerrit admin to CR and submit this, https://gerrit.wikimedia.org/r/#/c/432145/ [19:34:54] Project mwext-phpunit-coverage-publish build #4297: 04FAILURE in 26 sec: https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-publish/4297/ [19:38:18] (03CR) 10Jforrester: [C: 031] Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 (https://phabricator.wikimedia.org/T194242) (owner: 10MarcoAurelio) [19:56:22] twentyafterfour wondering could you press submit on https://gerrit.wikimedia.org/r/#/c/432145/ please? [19:57:25] paladox: done [19:57:32] twentyafterfour thanks :) [19:59:40] awight_mob https://gerrit.wikimedia.org/r/plugins/gitiles/scoring/ores/drafttopic/ works now! [19:59:40] awight_mob https://gerrit.wikimedia.org/r/plugins/gitiles/scoring/ores/drafttopic/ works now! [19:59:54] 10Release-Engineering-Team (Kanban), 10Release, 10Train Deployments: 1.32.0-wmf.3 deployment blockers - https://phabricator.wikimedia.org/T191049#4195130 (10mmodell) [20:00:20] 10Gerrit, 10Release-Engineering-Team, 10Scoring-platform-team (Current): Phabricator repo is not mirroring to gerrit as hoped - https://phabricator.wikimedia.org/T194295#4195134 (10Paladox) 05Open>03Resolved [20:10:44] 10Phabricator: Add support for task types - https://phabricator.wikimedia.org/T93499#4195156 (10mmodell) @Niedzielski Can you think of a better label? It should be unique to distinguish it from the regular task type. Maybe `"Create urgent task"` would be a little better? `"Create Task with a due date"` seems a... [20:12:12] 10Phabricator: Add support for task types - https://phabricator.wikimedia.org/T93499#4195159 (10greg) "urgent" might confuse people with "priority" (eg UBN!). :/ [20:13:20] 10Phabricator: Add support for task types - https://phabricator.wikimedia.org/T93499#4195162 (10mmodell) @greg: Good point. I'm really unsure maybe it should just be Create Task ... I really don't know if it should be a separate task type at all. [20:13:51] (03CR) 10Chad: [C: 032] Cleanup redundant rules [extensions/WantedPagesFromNS] (refs/meta/config) - 10https://gerrit.wikimedia.org/r/430056 (owner: 10MarcoAurelio) [20:16:11] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Different production servers have different versions of tidy installed, resulting in varying output - https://phabricator.wikimedia.org/T193414#4195174 (10Legoktm) [20:16:25] (03PS1) 10MaxSem: Add GlobalPreferences to extension-gate [integration/config] - 10https://gerrit.wikimedia.org/r/432165 (https://phabricator.wikimedia.org/T194229) [20:17:16] (03CR) 10MaxSem: [C: 04-1] "Waiting for the extension to be unbroken." [integration/config] - 10https://gerrit.wikimedia.org/r/432165 (https://phabricator.wikimedia.org/T194229) (owner: 10MaxSem) [20:17:50] (03CR) 10jerkins-bot: [V: 04-1] Add GlobalPreferences to extension-gate [integration/config] - 10https://gerrit.wikimedia.org/r/432165 (https://phabricator.wikimedia.org/T194229) (owner: 10MaxSem) [20:19:02] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4195192 (10Legoktm) [20:19:07] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4168635 (10Legoktm) [20:19:19] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4168635 (10Legoktm) 05duplicate>03declined Sorry, I duped the wrong way. [20:22:25] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4195214 (10zhuyifei1999) So what makes this declined? The merged task clearly di... [20:24:51] Yippee, build fixed! [20:24:52] Project mwext-phpunit-coverage-publish build #4298: 09FIXED in 2 min 26 sec: https://integration.wikimedia.org/ci/job/mwext-phpunit-coverage-publish/4298/ [20:29:16] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4195257 (10Legoktm) All servers are running stretch, so if there are still incon... [20:33:24] paladox: WIP: https://gerrit.wikimedia.org/r/#/c/432180/ [20:33:31] * paladox looks [20:33:39] oh! [20:34:20] yay [20:37:08] (03PS2) 10MaxSem: Add GlobalPreferences to extension-gate [integration/config] - 10https://gerrit.wikimedia.org/r/432165 (https://phabricator.wikimedia.org/T194229) [20:37:13] no_justification if we are planing to do 2.15 in a couple months or in the near future i can test this on gerrit-test3 :) [20:37:48] Near future [20:38:02] "Soon" :) [20:38:26] yay! [20:38:30] then i can begin testing [20:38:49] as i have already been testing it in other places but i mean in prod enviroment :) [20:39:38] (03CR) 10Umherirrender: [C: 031] Archive extension CategorySlideShow [integration/config] - 10https://gerrit.wikimedia.org/r/431610 (https://phabricator.wikimedia.org/T186969) (owner: 10MarcoAurelio) [20:40:10] (03CR) 10Umherirrender: [C: 031] Add tests for Extension:VEForAll [integration/config] - 10https://gerrit.wikimedia.org/r/431853 (owner: 10Cicalese) [20:40:25] (03Abandoned) 10Umherirrender: [VEForAll] Add unit tests [integration/config] - 10https://gerrit.wikimedia.org/r/431130 (owner: 10Umherirrender) [20:40:59] (03CR) 10Umherirrender: [C: 031] Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 (https://phabricator.wikimedia.org/T194242) (owner: 10MarcoAurelio) [20:42:36] no_justification i'll have to wait for it to be updated in https://gerrit.wikimedia.org/r/#/admin/projects/operations/software/gerrit [20:43:01] (under the 2.15 branch which also needs to be created :)) [20:47:55] (03CR) 1020after4: [C: 032] "looks good to me" [integration/config] - 10https://gerrit.wikimedia.org/r/415886 (https://phabricator.wikimedia.org/T177896) (owner: 10Mholloway) [20:49:52] (03Merged) 10jenkins-bot: Add mobileapps-diff-test periodic job [integration/config] - 10https://gerrit.wikimedia.org/r/415886 (https://phabricator.wikimedia.org/T177896) (owner: 10Mholloway) [21:14:28] 10Phabricator: Add support for task types - https://phabricator.wikimedia.org/T93499#4195385 (10MBinder_WMF) Love the discussion here. :) A few teams have adopted this task as their new default, because it offers everything a standard task does plus the Deadline option. Because of that, I like the idea of simpl... [21:15:47] twentyafterfour: hi, is it possible to merge and deploy https://gerrit.wikimedia.org/r/#/c/432150/ ? for CI thanks [21:16:12] (03CR) 1020after4: [C: 032] Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 (https://phabricator.wikimedia.org/T194242) (owner: 10MarcoAurelio) [21:16:22] Hauskatze: done [21:16:35] twentyafterfour: thanks sir :) [21:16:47] do we need to restart zuul to make it work? [21:16:59] no [21:17:00] no [21:17:03] lol [21:17:04] don't restart zuul [21:17:07] just reload it [21:17:22] well, restart, reload... I don't know the difference [21:17:28] there's a helper script for this, $ fab deploy_zuul [21:17:35] (03Merged) 10jenkins-bot: Revert "Archive Extension:AdvancedMeta" [integration/config] - 10https://gerrit.wikimedia.org/r/432150 (https://phabricator.wikimedia.org/T194242) (owner: 10MarcoAurelio) [21:18:54] do I have permission to do that? I'll try [21:19:09] if you have access to contint1001 [21:21:22] !log Reloading Zuul to deploy https://gerrit.wikimedia.org/r/#/c/432150/ [21:21:23] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [21:21:37] apparently I do [21:30:20] thanks twentyafterfour :) [21:30:24] * Hauskatze tests [21:31:17] it's testing on zuul [21:31:19] <3 [21:31:47] Hauskatze: you're welcome, glad to help [21:34:24] 10Continuous-Integration-Config, 10MediaWiki-extensions-AdvancedMeta, 10User-MarcoAurelio: Restore jenkins-bot normal tests (remove "archived") for extension AdvancedMeta - https://phabricator.wikimedia.org/T194242#4195440 (10MarcoAurelio) 05Open>03Resolved a:03MarcoAurelio Okay so the extension is una... [21:43:10] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4195475 (10ssastry) 05declined>03Open This has been fixed upstream https://g... [21:47:04] 10Release-Engineering-Team, 10MediaWiki-Parser, 10MediaWiki-Platform-Team, 10Operations, and 2 others: Servers using tidy-html5 are rendering pages differently, especially with - https://phabricator.wikimedia.org/T193414#4195487 (10ssastry) @zhuyifei1999 Tidy is being removed completely in 7 weeks ti... [21:48:17] (03PS2) 10Addshore: Add WikibaseLexeme to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/431305 (https://phabricator.wikimedia.org/T191463) [21:48:33] o/ ^^ that is ready to go now :) [21:52:04] (03PS2) 10MarcoAurelio: Whitelist Rxy on zuul [integration/config] - 10https://gerrit.wikimedia.org/r/432051 [21:58:16] 10Beta-Cluster-Infrastructure, 10User-Ryasmeen: Getting 503s on Beta cluster, Can't load VE - https://phabricator.wikimedia.org/T193795#4195542 (10Etonkovidova) All is fine - thanks, @Ryasmeen. [22:00:02] greg-g: I'd like to do an ORES deployment, if that sounds good? [22:03:54] 10Gerrit: mediawiki/extensions repo lacks extensions - https://phabricator.wikimedia.org/T194156#4195554 (10MarcoAurelio) >>! In T194156#4194894, @Umherirrender wrote: > Should be rewritten to use the rest api > > https://gerrit.wikimedia.org/r/projects/?d&format=json&p=mediawiki%2Fextensions%2F > Chad told me... [22:09:04] I'm getting a lot of this and don't understand why. Other submodules are fine: Check 'fetch_checks' failed: Synchronizing submodule url for 'submodules/articlequality' [22:09:29] legoktm hi, im wondering do you know how i can fix "The configuration setting 'wgAjaxExportList' was already set by another extension" please? [22:09:37] It dosen't tell me which extension to look at [22:11:03] awight: service or extension? I'll assume service whenever you ask, right? [22:11:23] yes, exactly. [22:12:32] It's usually a safe deployment since pages that rely on ORES degrade gracefully now... so I'm asking out of a show of etiquitte :p [22:13:52] twentyafterfour: I just noticed some horrific cruft: https://phabricator.wikimedia.org/source/ores-deploy/browse/master/scap/scripts/fetch_check.sh;c0db10224d57567cebebb8feb216f8a661b06019$9 [22:14:04] twentyafterfour: scap does all things submodule now, right? [22:14:45] awight: I think? [22:14:51] :) [22:15:04] I think it does a submodule sync indeed [22:15:40] 10Gerrit: mediawiki/extensions repo lacks extensions - https://phabricator.wikimedia.org/T194156#4195570 (10MarcoAurelio) I've added AdvancedMeta at https://gerrit.wikimedia.org/r/#/c/432298/ [22:15:48] Weird thing is happening, when I run "scap -s fetch" it fetches fine, but this fetch check crashes with an error from submodule sync. [22:16:21] awight: godspeed [22:16:25] the last thing I did was merge the gerrit-based submodule pointers, so I'm imagining this is somehow related to not rewriting URLs. [22:16:35] I'm removing the cruft to see if that helps. [22:19:40] twentyafterfour: This is interesting, https://phabricator.wikimedia.org/P7113 [22:20:37] wait, it's crazier than I thought. This should have been gerrit URLS [22:21:23] twentyafterfour: I think that "scap -s fetch" might be running * fetch * fetch_checks * fetch * fetch_checks [22:22:26] legoktm i think https://gerrit.wikimedia.org/r/#/c/374334/ should be reverted as it has the most un helpful error. [22:22:42] fwiw I really hate the url rewriting part of this process. it's ugly and fragile [22:22:48] * addshore has a question about beta, prod, extensions and extension-list [22:23:16] So, I just deployed WikibaseLexeme to beta, but forgot to add it to extension-list... [22:23:27] ... is it okay to add it to extension-list before it is in the make-wmf branch script? [22:23:43] twentyafterfour: I'm actually completely lost about what's happening. Gripping hair for a minute. [22:23:55] Check 'fetch_checks' failed: Already using interpreter /usr/bin/python3 [22:23:58] This kinda thing. [22:24:13] 10Gerrit: Please rewrite sync-with-gerrit.py to use Gerrit REST API - https://phabricator.wikimedia.org/T194318#4195590 (10MarcoAurelio) [22:24:18] all I'm trying to do is run "git lfs install" on all production machine by forcing scap fetch [22:24:38] addshore: not really, doesn't it need to be in the branch before it's in extension-list? [22:25:04] awight: let me see [22:25:14] I'll look at the log and see if I can see anything [22:25:54] scap/log/scap-sync-2018-05-09-0001.log [22:26:02] twentyafterfour: yes, I guess so. [22:26:29] * addshore pokes someone toward https://gerrit.wikimedia.org/r/#/c/431305/ (the make-wmf-branch patch) [22:26:42] Once that is merged I can add it to extension-list and fix i18n :) [22:28:28] (03CR) 1020after4: [C: 032] Add WikibaseLexeme to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/431305 (https://phabricator.wikimedia.org/T191463) (owner: 10Addshore) [22:28:36] Thanks! :) [22:28:47] awight: looks to me like the fialure is because of a dependency conflict with your wheels? [22:29:00] Double requirement given: mwreverts==0.1.5 from file:///srv/deployment/ores/deploy-cache/revs/2a099391e3363a07818d6887aee65f0b35fb5032/submodules/wheels/mwreverts-0.1.5-py2.py3-none-any.whl (already in mwreverts==0.0.6 from file:///srv/deployment/ores/deploy-cache/revs/2a099391e3363a07818d6887aee65f0b35fb5032/submodules/wheels/mwreverts-0.0.6-py2.py3-none-any.whl, name='mwreverts') [22:29:10] (03CR) 10Chad: [C: 032] Add WikibaseLexeme to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/431305 (https://phabricator.wikimedia.org/T191463) (owner: 10Addshore) [22:29:18] two different specific version numbers 0.1.5 vs 0.0.6 [22:29:40] 2+2@make-branch [22:29:45] :) [22:29:47] =4 [22:29:48] (03Merged) 10jenkins-bot: Add WikibaseLexeme to make-wmf-branch [tools/release] - 10https://gerrit.wikimedia.org/r/431305 (https://phabricator.wikimedia.org/T191463) (owner: 10Addshore) [22:29:50] =3 [22:29:54] <3 [22:29:54] <3 [22:29:57] twentyafterfour: ty, I'm cleaning that up now, but this has been crashing me too, > Check 'fetch_checks' failed: Synchronizing submodule url for 'submodules/articlequality' [22:30:41] awight: the log can be misleading, it might not be the exact line that folows failed: but it might be something on the following lines [22:31:04] because the failed check logger dumps more than one line, potentially, from the output of the failed command [22:31:19] kk I'm running with the other error fixed to know for sure [22:32:10] I wonder if it's problematic that we leave the changes to .gitmodules uncommitted [22:32:52] It's so nasty. I wish there were a transparent way to do this at the network or git level... [22:33:03] Thanks for providing me with counterspells, though :) [22:33:04] well there is actually [22:33:25] there is the insteadOf git config option [22:33:32] O_o [22:34:33] we should probably switch to using that, I think we avoided it because it requires changes to the gitconfig of the repo on the targets but I don't see that as a problem really right now [22:34:42] twentyafterfour: whoa. > Check 'fetch_checks' failed: Already using interpreter /usr/bin/python3 [22:35:04] that specific logline was appearing but not causing a crash, previous. [22:35:47] awight: look further down, it's another dependency conflict [22:35:50] * awight blinks [22:35:51] hargh [22:36:08] it logs the entire output from the failed check, not just the failed command [22:36:23] so any check that fails starts with non-failing noise in the log [22:36:44] right, I see what you're saying [22:37:19] because scap doesn't run the check script one line at a time and capture the output from each, it just runs the whole script and captures all the output from that [22:37:41] but the deploy-log kindof obscures that fact [22:38:19] because there is no delineation from the "Check failed:" message and the beginning of the output from the script starting on that same line [22:38:40] (03PS1) 10Mholloway: Run mobileapps-periodic-test hourly regardless of repo changes [integration/config] - 10https://gerrit.wikimedia.org/r/432310 (https://phabricator.wikimedia.org/T177896) [22:39:06] thcipriani: no_justification: do you see any reason why we shouldn't convert the git url rewriting to use insteadOf in the config file rather than manually editing .gitmodules? [22:39:15] it seems a lot cleaner to me [22:39:25] That makes a lot of sense [22:40:12] [url ""] [22:40:12] insteadOf = https://gerrit..... [22:40:28] actually not manual editing but automated kludge is a kludge nonetheless [22:40:30] (03PS2) 10Mholloway: Run mobileapps-periodic-test hourly regardless of repo changes [integration/config] - 10https://gerrit.wikimedia.org/r/432310 (https://phabricator.wikimedia.org/T177896) [22:41:41] no_justification: yeah that exactly. We'd just have to add that rule to the deploy user on the target machines [22:42:02] the deploy user's ~/.gitconfig [22:42:49] now that doesn't cover situations where gerrit is not the upstream [22:43:18] we could have a second rule for phab, and I guess we could cover github/bitbucket if those are used for any deployments [22:43:38] Easy enough though [22:44:02] You could also do it when initially cloning the repo instead of ~/.gitconfig [22:44:04] that url rewriting code in scap has caused a lot of headaches and it feels really fragile to me [22:44:09] Do a `git config` on the actual repo [22:44:09] twentyafterfour: lol it was the mundane duplicate wheel error, solved now. [22:44:19] no_justification: yeah [22:45:09] awight: I'm not sure how to make that more clear. I guess I could change the log message to say "Output of the failed script:\n" [22:45:36] twentyafterfour: IMO I'd prefer to get the entire script output inlined in the deploy-log [22:45:37] could also prefix each line of the failed script output with "CHECK:" or something [22:45:52] awight: it should be already doing that [22:46:31] I think it only logs if the check fails though [22:46:32] twentyafterfour: OOOH thanks [22:46:47] that's right, I remember you explaining this to me a month ago /o\ [22:47:15] from what I remember, the git module is not on a place on the deployment server where we can just rewrite the url [22:47:21] you should see both stdout and stderr from the check script, merged into the deploy log, afaik [22:48:38] that is, a submodule on the deployment server is located in the /srv/deployment/[reponame]/.git/modules/[submodule-path] directory [22:49:10] legoktm wondering would using callback work around the problem? [22:49:19] with wgAjaxExportList? [22:49:37] /srv/mediawiki/w/extensions/MsCalendar/extension.json:66: "AjaxExportList": [ [22:49:42] so it's the submodule path + the url that has to be updated [22:49:46] thcipriani: the path prefix can be altered by the rewrite but I'm not sure how we'd handle the suffix [22:49:51] /srv/mediawiki/w/extensions/AJAXPoll/extension.json:67: "AjaxExportList": [ [22:50:02] /srv/mediawiki/w/extensions/MsUpload/extension.json:57: "AjaxExportList": [ [22:50:20] because it looks like we rewrite something.git to .git/modules/something [22:50:30] indeed [22:51:23] and that's easy to rewrite from the submodule path in a repo, but that might not have any relation to its location in gerrit [22:51:34] so it's tricky :\ [22:52:34] yeah that sucks, I thought I was onto something [22:54:37] it seems like it'd be cleaner if we didn't modify .gitmodules but instead just touch the .git/config and leave .gitmodules alone [22:55:11] because leaving the target repo with unstaged/uncommitted changes feels dirty [22:55:24] we could checkout .gitmodules after the submodule sync that modifies .git/config [22:55:30] yeah [22:55:47] that would still be a dirty kludge but it would be slightly better, imo [22:56:11] but would there be problems with subsequent syncs? I don't know [22:56:29] yeah if people do git submodule sync in their check scripts like ores does ;) [23:02:01] legoktm or twentyafterfour or reedy could you review https://gerrit.wikimedia.org/r/#/c/432316/ please? :) [23:03:55] (We stopped doing the redundant submodule stuff as of today :) [23:04:16] paladox: +2 [23:04:23] twentyafterfour thanks :), could you also do https://gerrit.wikimedia.org/r/#/c/432317/ please? :) [23:04:42] (will backport to REL1_31 once merged) [23:05:00] ffs beta hurry up .... https://integration.wikimedia.org/ci/job/beta-scap-eqiad/207098/console [23:05:03] thanks :) [23:05:03] paladox: done [23:05:06] :) [23:06:15] postmerge looks stuck for parsoid [23:06:16] https://integration.wikimedia.org/zuul/ [23:06:18] 7hrs [23:07:39] thcipriani ^^ [23:09:11] addshore: that is well outside the normal time for that job :-/ [23:09:39] i wonder if i should go kill it [23:10:06] occasionally it takes a while [23:10:20] https://integration.wikimedia.org/ci/blue/organizations/jenkins/beta-scap-eqiad/detail/beta-scap-eqiad/207074/pipeline/ took 45 mins [23:12:07] twentyafterfour addshore it could be hhvm [23:12:13] it appears to be running still not locked up [23:13:18] I guess it only takes a long time when there are actual changes to messages that require regenerating the l10n cache? [23:14:14] there is a pretty nasty backlog in zuul, not just the one stalled for 7 hours [23:14:53] it seems it [23:15:24] it seems it's quibble [23:15:28] hrm, the one stalled for 7 hours may just be stuck somewhere in the queue after I upgraded jenkins install [23:15:49] paladox: why quibble [23:16:00] twentyafterfour it's on most mw extensions [23:16:03] and runs tests [23:16:08] the 45 minute beta-scap-eqiad job is probably https://phabricator.wikimedia.org/T191921 [23:16:09] https://integration.wikimedia.org/ci/job/quibble-vendor-mysql-hhvm-docker/1690/console [23:16:21] thcipriani: yeah [23:16:56] there were other doc update jobs that ran for parsoid after that one, so that's good. Not exactly sure how to clear it out of the queue though :\ [23:17:24] finally syncing :) [23:21:07] twentyafterfour could you merge https://gerrit.wikimedia.org/r/#/c/432323/ and https://gerrit.wikimedia.org/r/#/c/432324/ pelase? :) [23:21:10] (backports) [23:22:07] twentyafterfour thanks :) [23:22:53] thcipriani it seems the other jobs are getting stuck in postmerge [23:24:31] they're all waiting on beta-scap-eqiad I'd bet [23:24:36] I'll keep an eye on it [23:26:25] ok thanks [23:36:34] twentyafterfour: it was not appropriate to merge those patches...clearly no one tested them [23:37:20] legoktm though it was done here https://github.com/wikimedia/mediawiki-extensions-MsUpload/commit/7fb7b2c77ac214295515b5d2b7b917696f5be201 [23:37:31] 3 years old [23:38:03] https://gerrit.wikimedia.org/r/#/c/258707/ [23:38:21] clearly no one tested that one either!! [23:38:55] legoktm well the user did v+2 [23:39:29] Yes, that means they clicked it [23:39:33] Doesn't mean they tested it [23:39:49] legoktm it's also done here https://github.com/wikimedia/mediawiki-extensions-TopLists/commit/3aec3b3f6dcf98b1eeaac79024538e3bc7299987 [23:39:58] https://github.com/wikimedia/mediawiki-extensions-TopLists/blob/442ce1ea6f8d81a2b5f5e25f91a39d0c78e74d91/extension.json#L48 [23:40:23] great they're all broken [23:40:32] huh? [23:40:38] how does one use that var then? [23:40:47] you don't! [23:40:48] https://phabricator.wikimedia.org/T122878 [23:40:54] it's deprecated and long should be dead [23:41:32] legoktm any example commits on migrating to api? [23:42:21] not off the top of my head [23:42:57] * legoktm -> afk, class [23:48:27] ok [23:48:30] thanks [23:55:09] paladox: so we need to revert those then? [23:55:19] twentyafterfour i doin't think so [23:55:23] as it would not work before [23:55:29] and will break with 1.31 anyways [23:55:34] hmm [23:56:02] so what to do then? [23:56:24] Not sure apart from converting it to use the api instead (which im not sure what to do) [23:56:38] but it's been broken for a while it seems and im not sure how no one noticed. [23:57:33] 10Phabricator: Add support for task types - https://phabricator.wikimedia.org/T93499#4195862 (10mmodell) ok I changed it to just say "Create Task" in the form title. I'm also looking into the possibility of hooking into the calendar system to implement deadline notifications (via email / phabricator built-in web... [23:58:45] Crap extensions no one actually uses [23:58:52] Or cares about [23:58:56] Or they don't use that feature in it