[00:00:34] special version claims beta is at https://gerrit.wikimedia.org/r/plugins/gitiles/mediawiki/extensions/CentralNotice/+log/1530ef096904761dda2d2213c3b2d726f9b009d2/ [00:00:40] https://gerrit.wikimedia.org/r/plugins/gitiles/mediawiki/extensions/CentralNotice/+log/refs/heads/master [00:00:45] https://gerrit.wikimedia.org/r/plugins/gitiles/mediawiki/extensions/CentralNotice/+log/refs/heads/wmf_deploy [00:00:53] 5 and 7 weeks ago != 2 months ago [00:01:25] aaargh [00:02:11] Krinkle: Reedy: for me on Special:Verison I don't see that commit https://en.wikipedia.org/wiki/Special:Version [00:02:20] but rather https://gerrit.wikimedia.org/g/mediawiki/extensions/CentralNotice/+/a8ee3ec72388b27dfe5c72ad9bb7dc3cd074c42a [00:02:45] which is one of the parents of the commits I get on origin/wmf_deploy [00:03:10] which doesn't surprise me too much because for a while I have the impression that Special:Version has been skipping merge commits [00:04:03] AndyRussG: beta cluster [00:04:24] I'm talking about beta [00:04:34] ahhh oops [00:04:35] https://en.wikipedia.beta.wmflabs.org/wiki/Special:Version [00:04:55] because I suspect it's actually tracking master instead of wmf_deploy [00:04:59] right heheh that's what happens when you reply to a ping word but don't bother to read backscroll heh sorry [00:05:05] yes beta cluster should track master [00:05:09] thus extra confusing / lowering confidence [00:05:13] well, it shoudl track what goes to prod soon [00:05:20] which for CN != master [00:05:39] hmmm [00:05:41] CN@master is just a more difficult way of managing commits that have +1s on them that have't been merged/deployed yet. [00:05:48] which is what I'm suggesting we do instead [00:05:59] well we do use the beta cluster for testing CN stuff before it goes to prod [00:06:04] people can work on CN just as well and +1 them for you to +2 whenver its good and ready [00:06:59] right, except prod can completely break and be down with 500 errors n every page view when beta cluster is fine today [00:07:06] ... if we removed a method used in CN@wmf_deploy [00:07:13] and merged a fix for it in master ahead of time [00:07:49] normally the workflow is everything is only merged on master first [00:08:10] then we merge master into wmf_deploy, or cherry pick from master if urgent [00:08:31] oh I see what you mean [00:08:42] argh [00:09:01] yeah, but thats undiscoverable and seems to not have much benefit othe than that it may have seemed easier to set up this way at a time where we didn't know how to restrict gerrit +2 permissions. [00:10:00] that is, unless it becomes a weekly event for you to sync master into wmf_deploy and to block the train unless that has happened. [00:11:04] K right, so in essence, the distance between what's deployed for CN vs what's deployed for core is much greater than the equivalent distance between the two on the beta cluster [00:11:21] yeah let's make a task [00:14:35] (also scratch what I said above about prod Special:Version not being quite right... argh getting late, coffee effect has worn off and it's too late for more) [00:16:00] also as you may or may not have heard, there has been talk for some time now about CentralNotice being handed off and no longer maintained by FR-Tech, though I have no idea if it'll really happen one day [00:16:20] commented at https://phabricator.wikimedia.org/T179536 [00:17:21] AndyRussG: ^ [00:17:33] Krinkle: yep thanks much! [00:17:36] I think the Special:Version issue on beta is just a caching problem with the git hash [00:17:49] from what I can tell on the server, it has actually deployed master - unfortunately [00:17:54] (instead of wmf_deploy) [00:18:13] which makes sense since the exception we crafted for production's wmf branching source was afaik never applied to beta [00:20:09] right [00:21:13] Krinkle thanks for the clear explanation in the comment, that's really helpful! Also thx Reedy :) [00:22:53] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.36-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Reedy) [00:24:49] AndyRussG: I hope the new model allows you(r team) to more easily apply some matches more piece-meal [00:24:58] e.g. unimportant ones can jsut go out as you spot them. [00:25:14] not saying you'll have to spend more time on this, but might be win-win if it's not under the burden of "gotta do a big merge" [00:25:36] yep [00:27:03] (speaking of which, I should actually do one of those this evening before the branch cut) [00:30:04] deployment-mediawiki11:/srv/mediawiki/php-master/cache/gitinfo$ cat info-extensions-CentralNotice.json [00:30:04] {"head": "31804f3490331b9d168d239f5ba1517ac0c8a2a6", [00:30:10] * Krinkle continues debugging beta [00:30:17] well it loks likek the git info cache is fine [00:30:21] that's master's head [00:32:18] {"head": "31804f3490331b9d168d239f5ba1517ac0c8a2a6", "remoteURL": "https://gerrit.wikimedia.org/r/mediawiki/extensions/CentralNotice", "branch": "31804f3490331b9d168d239f5ba1517ac0c8a2a6", "headCommitDate": "1611900309", "headSHA1": "1530ef096904761dda2d2213c3b2d726f9b009d2", "@directory": "/srv/mediawiki-staging/php-master/extensions/CentralNotice"}krinkle@deployment-mediawiki11:/srv/mediawiki/php-master/cache/gitinfo$ [00:32:21] oh but there's more [00:32:28] headSHA1 != head [00:32:40] great data model this is [00:34:41] aparently 'head' is supposed to be the sha1 (if detached) or the ref (e.g. refs/head/master) [00:34:48] and 'headSHA1' resolved to definitely a sha1 [00:34:58] except... they are both sha in this case, and... different [00:35:37] hmmm fun 8p [00:37:17] ah but scap has its own re-creation of this [00:37:25] so this isn't running https://github.com/wikimedia/mediawiki/blob/master/includes/GitInfo.php [00:38:12] but https://github.com/wikimedia/scap/blob/48fe406ebebdfc77c3fe2f94238e8e0fc87144a6/scap/git.py [00:41:20] hmmm [00:45:11] > -bash: cd: .git: Not a directory [00:45:18] [00:44 UTC] krinkle at deployment-deploy01.deployment-prep.eqiad.wmflabs in /srv/mediawiki-staging/php-master/extensions/CentralNotice ((HEAD detached at 31804f34)) [00:45:18] 1s $ cat .git [00:45:18] gitdir: ../.git/modules/CentralNotice [00:45:21] Of course. [00:45:35] OK, I guess that's what Reedy meant earrlier by mediawiki-extensions messing it up [00:47:27] > git_dir = resolve_gitdir(directory) [00:47:28] gut [00:47:29] but [00:47:32] head_sha1 = get_disclosable_head(directory, branch) [00:48:57] php-master/extensions/CentralNotice$ /usr/bin/git rev-list -1 @{upstream} [00:48:57] > fatal: HEAD does not point to a branch [00:49:12] remotes/origin/HEAD -> origin/wmf_deploy [00:49:14] interesting [00:49:26] so the "default branch" for this repo actually does point to wmf_deploy [00:49:46] which e.g. affects what `git clone` returns by default, and repo browsers [00:50:09] and then it falls back to finding a merge base [00:50:26] because it thinks all master commits since the last merge are e.g. security patches or other non-public things [00:50:43] because usually when HEAD isn't found in the remote that's the reason [00:50:44] xD [00:51:12] "1614500 Merge branch 'master' into wmf_deploy by Andrew Green · 2 months ago" [00:51:20] yeah, thats the last time the two branches had a common merge point [00:51:28] so it shows the last non-merge commit before that [00:51:32] weee [00:52:16] hmmm [00:52:44] K well also sounds like an issue that would go away by harmonizing deploy procedure [00:53:22] Krinkle were there any specific CN issues you were looking into when you came across this? I am actually about to merge CN master into wmf_deploy.. [00:53:43] AndyRussG: nope, just general confusion that reminded me when I noticed this [00:54:21] I found that CN task in an unrelated search and tried to figure out what its state is in beta [00:56:53] ahh okok welp sounds like a point in favour of states of general confusion then [00:57:13] I did also look for other tasks about CN deploy pain points, but it seemed all others have since been closed [01:20:09] 10phan, 10MediaWiki-Core-Testing: Phan failed to notice second arg passed to a method that only takes one - https://phabricator.wikimedia.org/T279976 (10Krinkle) [01:44:17] Krinkle for ^ the function given in the description is serialize, but the method call is unserialize - should the function be unserialize? [01:54:21] 10phan, 10MediaWiki-Core-Testing: Phan failed to notice second arg passed to a method that only takes one - https://phabricator.wikimedia.org/T279976 (10Krinkle) [01:54:24] DannyS712: applies to both actually [02:29:24] 10Continuous-Integration-Infrastructure, 10Composer: Re-build CI containers with Composer 2.0 - https://phabricator.wikimedia.org/T279857 (10Reedy) [05:23:41] James_F when I have a chance, sent the next two patches to merge container.php entries - https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Flow/+/678648 and follow up in the relation chain [05:23:55] *when you have a chance [05:45:33] Krinkle: Reedy: beta has three deploy hosts because I'm migrating them to buster (01 and 02 are stretch, 03 is buster) but that project is blocked on Jenkins agent puppet classes not liking some new cloudvps features that the new deploy host is using [06:20:54] Majavah: is one of the stretch ones redundant/unused? Or do we aim to always have a spare? [06:42:43] Krinkle: I guess 02, since 03 is otherwise fully functional but it just can't be used for the jenkins automated deployments without some hacks yet [06:45:43] I don't think we always need a spare, but I guess it's nice to have since setting one up without having a working one can be a pain [07:33:39] 17:49:26 so the "default branch" for this repo actually does point to wmf_deploy <-- which has been really annoying :( [07:55:20] 10Release-Engineering-Team (Deployment services), 10Release-Engineering-Team-TODO (2020-10-01 to 2020-12-31 (Q2)), 10Scap, 10Patch-For-Review, 10User-jijiki: Update Scap to perform rolling restart for all MW deploy - https://phabricator.wikimedia.org/T266055 (10jijiki) [08:13:47] 10Release-Engineering-Team-TODO (2021-04-01 to 2021-06-30 (Q4)), 10Release, 10Train Deployments: 1.37.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T278345 (10Urbanecm) Hello, is this going to be wmf.1 deployment, or wmf.39 deployment? https://gerrit.wikimedia.org/r/c/mediawiki/core/+/67867... [08:14:52] 10Gerrit: Shouldn't be possible to push direct to master - https://phabricator.wikimedia.org/T279989 (10Samwilson) [08:37:42] 10Gerrit: Shouldn't be possible to push direct to master - https://phabricator.wikimedia.org/T279989 (10Paladox) https://gerrit.wikimedia.org/r/admin/repos/mediawiki/extensions/MsUpload,access Shows it has push rules. Should be as simple as removing that rule to revoke direct pushing. [09:54:07] anyone around who can help with a Zuul / gate-and-submit issue? [09:54:10] https://gerrit.wikimedia.org/r/c/mediawiki/extensions/WikibaseCirrusSearch/+/678623 should be ready to go [09:54:17] but it’s not appearing in gate-and-submit [09:54:38] and I’m guessing that’s because the Change-Id of the Wikibase change it depends on was accidentally also used for the WikibaseCirrusSearch change at first [09:55:00] so now there’s an abandoned WikibaseCirrusSearch change with the same Change-Id that the real WikibaseCirrusSearch change depends on :/ [09:55:14] and the Wikibase change with that Change-Id was already merged, so we can’t change that anymore [09:55:30] what’s the best way forward? [09:55:36] Lucas_WMDE: given the patch that you want it to depend on is merged, can't you just remove the depends on line? [09:56:01] I guess that would be one option [09:56:08] but might cause confusion when backporting the change… [09:57:30] it looks like manually submitting the change would be another option, and I’m not sure if that would be worse [09:59:55] 10phan, 10MediaWiki-Core-Testing: Phan failed to notice second arg passed to a method that only takes one - https://phabricator.wikimedia.org/T279976 (10Daimona) The first line of that file reads > // phpcs:ignoreFile -- It's an external lib and it isn't. Let's not bother. And as such, this file is also excl... [10:00:58] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.36-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Daimona) [10:01:47] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Epic, 10MW-1.36-release: Expand the set of bundled extensions and skins in MediaWiki 1.36 - https://phabricator.wikimedia.org/T246381 (10Daimona) >>! In T246381#6992559, @Jdforrester-WMF wrote: > I think it's reasonable to pu... [10:33:45] uploaded a new patch set to turn Depends-On into X-Depends-On, let’s see if that helps [12:22:10] (03PS1) 10Reedy: Add gpeterson to CI allow list [integration/config] - 10https://gerrit.wikimedia.org/r/678829 (https://phabricator.wikimedia.org/T280004) [12:23:24] (03PS1) 10Reedy: Change 'Trusted long term users' to 'Trusted users' [integration/config] - 10https://gerrit.wikimedia.org/r/678830 [12:23:37] (03CR) 10Reedy: [C: 03+2] Add gpeterson to CI allow list [integration/config] - 10https://gerrit.wikimedia.org/r/678829 (https://phabricator.wikimedia.org/T280004) (owner: 10Reedy) [12:23:56] (03CR) 10Reedy: [C: 03+2] Change 'Trusted long term users' to 'Trusted users' [integration/config] - 10https://gerrit.wikimedia.org/r/678830 (owner: 10Reedy) [12:24:59] (03CR) 10Reedy: [C: 03+2] "Though, maybe it should be "Trusted Contributors" like https://gerrit.wikimedia.org/r/admin/groups/2021f25e7515187a81d51f8fe14dd6f25617cce" [integration/config] - 10https://gerrit.wikimedia.org/r/678830 (owner: 10Reedy) [12:25:52] (03Merged) 10jenkins-bot: Add gpeterson to CI allow list [integration/config] - 10https://gerrit.wikimedia.org/r/678829 (https://phabricator.wikimedia.org/T280004) (owner: 10Reedy) [12:26:07] (03Merged) 10jenkins-bot: Change 'Trusted long term users' to 'Trusted users' [integration/config] - 10https://gerrit.wikimedia.org/r/678830 (owner: 10Reedy) [12:26:47] !log Reloading Zuul to deploy https://gerrit.wikimedia.org/r/678829 https://gerrit.wikimedia.org/r/678830 T280004 [12:26:50] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [12:26:50] T280004: Request "Add Patch Set" permission on Kartographer-extension for gpeterson user - https://phabricator.wikimedia.org/T280004 [13:03:12] 10Release-Engineering-Team (Deployment services), 10Release-Engineering-Team-TODO (2021-04-01 to 2021-06-30 (Q4)), 10Scap, 10SRE, 10Python3-Porting: Porting scap to Python 3 - https://phabricator.wikimedia.org/T279628 (10MoritzMuehlenhoff) p:05Triage→03High [13:18:35] 10Release-Engineering-Team-TODO (2021-04-01 to 2021-06-30 (Q4)), 10Release, 10Train Deployments: 1.37.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T278345 (10thcipriani) >>! In T278345#6993194, @Urbanecm wrote: > Hello, is this going to be wmf.1 deployment, or wmf.39 deployment? https://ge... [13:37:31] 10Release-Engineering-Team (Development services), 10Gerrit (Gerrit 3.3): Switch Gerrit from Java 8 to Java 11 - https://phabricator.wikimedia.org/T268225 (10thcipriani) >>! In T268225#6992220, @Paladox wrote: > I think that we use java 11 now? Looking at ps output on gerrit1001: we're still using java 8 >... [14:42:15] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.36-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Jdforrester-WMF) [Moving specific discussion here.] >>! In T246381#6993548, @Daimona wrote: >>>! In T246381#6992559, @Jdfo... [14:42:17] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Epic, 10MW-1.37-release: Expand the set of bundled extensions and skins in MediaWiki 1.37 - https://phabricator.wikimedia.org/T279842 (10Jdforrester-WMF) [14:42:20] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Epic, 10MW-1.36-release: Expand the set of bundled extensions and skins in MediaWiki 1.36 - https://phabricator.wikimedia.org/T246381 (10Jdforrester-WMF) [14:42:23] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.36-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Jdforrester-WMF) [14:42:28] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.36-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Jdforrester-WMF) [14:42:46] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Tracking-Neverending: Expand the set of bundled extensions and skins to achieve a default MediaWiki experience that's comparable to Wikimedia sites - https://phabricator.wikimedia.org/T178349 (10Jdforrester-WMF) [14:42:48] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Epic, 10MW-1.36-release: Expand the set of bundled extensions and skins in MediaWiki 1.36 - https://phabricator.wikimedia.org/T246381 (10Jdforrester-WMF) 05Open→03Resolved a:03Jdforrester-WMF Nothing to do. [14:43:07] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.37-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Jdforrester-WMF) [15:19:29] 10Phabricator, 10Security-Team: Enable risk rating field in Phabricator's task form - https://phabricator.wikimedia.org/T279690 (10sbassett) p:05Triage→03Medium [15:27:36] 10Gerrit: Shouldn't be possible to push direct to master - https://phabricator.wikimedia.org/T279989 (10Legoktm) 05Open→03Resolved a:03Legoktm >>! In T279989#6993288, @Paladox wrote: > https://gerrit.wikimedia.org/r/admin/repos/mediawiki/extensions/MsUpload,access Shows it has push rules. Should be as simp... [15:34:10] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Epic, 10MW-1.36-release: Expand the set of bundled extensions and skins in MediaWiki 1.36 - https://phabricator.wikimedia.org/T246381 (10DannyS712) 05Resolved→03Declined Per precedent at T196650 - no extensions were actua... [15:34:12] 10MediaWiki-Releasing, 10MediaWiki-Installer, 10MediaWiki-Stakeholders-Group, 10Tracking-Neverending: Expand the set of bundled extensions and skins to achieve a default MediaWiki experience that's comparable to Wikimedia sites - https://phabricator.wikimedia.org/T178349 (10DannyS712) [15:41:01] 10Release-Engineering-Team (Logspam), 10Structured-Data-Backlog (Current Work), 10Wikimedia-production-error: Cannot use 'in' operator to search for 'filemime' in undefined at getMediaFilters - https://phabricator.wikimedia.org/T279393 (10CBogen) [15:43:29] 10MediaWiki-Releasing, 10AbuseFilter (Overhaul-2020), 10MW-1.37-release, 10Patch-For-Review: Bundle AbuseFilter extension with MediaWiki - https://phabricator.wikimedia.org/T191740 (10Daimona) >>! In T191740#6994610, @Jdforrester-WMF wrote: > Sorry, I was insufficiently clear. Support for postgres and sqli... [15:54:58] (03PS1) 10Jforrester: deployments-calendar: MediaWiki, not Mediawiki [tools/release] - 10https://gerrit.wikimedia.org/r/678880 [15:55:06] legoktm: https://gerrit.wikimedia.org/r/c/mediawiki/tools/release/+/678880 :-) [15:55:28] ty :D [15:55:32] (03CR) 10Legoktm: [C: 03+2] "Thanks :D" [tools/release] - 10https://gerrit.wikimedia.org/r/678880 (owner: 10Jforrester) [15:55:44] Any time. [16:01:48] (03Merged) 10jenkins-bot: deployments-calendar: MediaWiki, not Mediawiki [tools/release] - 10https://gerrit.wikimedia.org/r/678880 (owner: 10Jforrester) [16:22:25] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Cloud-VPS, 10Patch-For-Review, 10cloud-services-team (Kanban): Support Cinder for CI docker workers - https://phabricator.wikimedia.org/T277078 (10Andrew) p:05Triage→03Medium [16:35:20] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO (2020-10-01 to 2020-12-31 (Q2)), 10Quibble, and 2 others: Terminating MySQL takes several minutes in (Wikibase?) CI jobs - https://phabricator.wikimedia.org/T265615 (10Andrew) [16:36:06] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services), 10Cloud-VPS, 10Patch-For-Review, 10cloud-services-team (Kanban): integration instances suffer from high IO latency due to Ceph - https://phabricator.wikimedia.org/T266777 (10Andrew) 05Open→03Resolved a:03And... [16:41:35] !log deleting errant wmf/1.36.0-wmf.39 branches in mediawiki/core and submodule repos [16:41:37] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [16:43:18] marxarelli: Looks good to me, thanks! [16:43:33] James_F: still running :) [16:43:44] the git push --delete is a bit slow [16:43:53] Removed from core and the VE sub-module; I assumed those would be first and last respectively. [16:43:56] * James_F shrugs. [16:44:42] ftr i did `gerrit ls-projects --show-branch wmf/1.36.0-wmf.39 | while read _ repo; do git push ssh://dduvall@gerrit.wikimedia.org:29418/$repo --delete wmf/1.36.0-wmf.39; done` [16:45:17] not `git submodule foreach`, so that's probably why the order is different [16:45:17] Fancy. [16:45:22] But right. [16:46:15] ok, all done! [16:59:47] 10Continuous-Integration-Config, 10Better Use Of Data, 10Product-Data-Infrastructure: Add CI jobs for mediawiki/libs/metrics-platform - https://phabricator.wikimedia.org/T279180 (10Mholloway) [17:02:04] is it just me or does npm no longer have a "report security issue" button on each package page? [17:03:06] Hey folks. I had an issue doing a beta deployment for ORES. Looks like I couldn't connect to the deployment host (deployment-ores01) from deployment-deploy01. See https://phabricator.wikimedia.org/T278723#6995430 for the full error. [17:03:20] It's been a long time since the last ORES deployment so I imagine something might have changed. [17:03:24] Like 9 months. [17:05:49] hmmm, deployment-deploy01:/etc/ssh/ssh_known_hosts and deployment-ores01:/etc/ssh/ssh_host_ecdsa_key.pub have the same key [17:07:49] 10LibUp: Update libup for Gerrit's new ssh host keys - https://phabricator.wikimedia.org/T257382 (10Dzahn) [17:19:51] I opened a npm support ticket asking how to get advisories issued [18:02:06] halAFK, Majavah, my drive-by comment is that maybe you need to update your targets to use the wikimedia.cloud domain rather than .wmflabs [18:02:25] especially if any of the concerned hosts have been rebuilt since the great rename [18:06:15] 10Beta-Cluster-Infrastructure, 10Beta-Cluster-reproducible, 10User-Majavah, 10User-zeljkofilipin: Cannot login to beta cluster: "There seems to be a problem with your login session..." - https://phabricator.wikimedia.org/T263617 (10Ryasmeen) 05Resolved→03Open Re-opening this task since I am having this... [18:07:06] Is there any CI pattern people use to test integration between a MediaWiki install and a backend service? I know quibble obviously doesn't run our docker-compose bespoke environment… [18:10:18] 10Beta-Cluster-Infrastructure, 10Beta-Cluster-reproducible, 10User-Majavah, 10User-zeljkofilipin: Cannot login to beta cluster: "There seems to be a problem with your login session..." - https://phabricator.wikimedia.org/T263617 (10Majavah) `lang=shell-session taavi@deployment-sessionstore03:~$ systemctl s... [18:19:38] James_F: vague sense that the docker-compose env was getting used for... something [18:19:42] (in CI) [18:23:40] 10Continuous-Integration-Infrastructure: Post-merge builds failing with "Command 'mwext-phpunit-coverage' returned non-zero exit status 1" - https://phabricator.wikimedia.org/T280064 (10Tgr) [18:43:20] 10Beta-Cluster-Infrastructure, 10Beta-Cluster-reproducible, 10User-Majavah, 10User-zeljkofilipin: Cannot login to beta cluster: "There seems to be a problem with your login session..." - https://phabricator.wikimedia.org/T263617 (10Ryasmeen) >>! In T263617#6995736, @Majavah wrote: > `lang=shell-session > t... [18:58:49] brennen: Hmm. Maybe something experimental? [18:59:52] brennen: No hits for `compose[^r]` in jjb, so… [19:00:02] yeah, was just looking [19:00:34] Our use case is the testing the integration of the WikiLambda extension with the function-orchestrator service. [19:00:53] * James_F had foolishly assumed that MW<->services was a thing we had CI testing for. [19:01:44] 10Release-Engineering-Team (Development services), 10Gerrit (Gerrit 3.3): Switch Gerrit from Java 8 to Java 11 - https://phabricator.wikimedia.org/T268225 (10hashar) The puppet bits are preparation work to let us "trivially" switch from java 8 to java 11. What worries me is whether our compiled plugin would a... [19:01:47] brennen: Should I make a "please make it possible" epic task for y'all? :-) [19:03:20] CirrusSearch uses the cindy-browser-test-bot [19:03:33] Does that use docker-compose? [19:03:52] I thought it was old enough that it used vagrant or something similarly terrible. [19:04:13] no idea [19:04:21] but that's the only MW<->services testing I'm aware of [19:04:28] Ack. [19:04:39] With m8s obviously this will be fixed soon, right? ;-) [19:13:21] 10Gerrit, 10Release-Engineering-Team (Development services): Clean GerritSite.css from GWT styling - https://phabricator.wikimedia.org/T277645 (10hashar) 05Open→03Resolved a:03Paladox Thank you @Paladox ! [19:16:18] 10Release-Engineering-Team (Deployment services), 10Release-Engineering-Team-TODO (2021-04-01 to 2021-06-30 (Q4)): Automate the merge of the weekly branch cut - https://phabricator.wikimedia.org/T278993 (10dduvall) 05Open→03Resolved [19:16:21] 10Release-Engineering-Team (Deployment services), 10Release-Engineering-Team-TODO, 10Scap, 10Epic, 10Goal: Automate the Train - https://phabricator.wikimedia.org/T196515 (10dduvall) [19:25:11] (03PS1) 10Ahmon Dancy: jjb: Rename beta-scap-eqiad to beta-scap-sync-world [integration/config] - 10https://gerrit.wikimedia.org/r/678927 [19:29:49] 10Release-Engineering-Team (Deployment services), 10Release-Engineering-Team-TODO (Yak Shaving 🐃🪒 ): Automatically tag the release task in the commit message of the weekly branch cut - https://phabricator.wikimedia.org/T278994 (10Majavah) > We'd need a way to find out what the next branch task is; https://trai... [19:31:58] James_F: cindy still uess vagrant. Some day i will finally find the time to make quibble install a cluster of wikis or something... [19:32:06] Ack. [19:43:24] 10Continuous-Integration-Infrastructure, 10Release-Engineering-Team (CI & Testing services): Provide a CI mechanism by which we can test end-to-end integration between MW and back-end services, e.g. via docker-compose - https://phabricator.wikimedia.org/T280070 (10Jdforrester-WMF) [19:46:12] (03CR) 10Jforrester: "LGTM. Want me to deploy?" [integration/config] - 10https://gerrit.wikimedia.org/r/678927 (owner: 10Ahmon Dancy) [19:47:07] (03CR) 10Ahmon Dancy: "> Patch Set 1:" [integration/config] - 10https://gerrit.wikimedia.org/r/678927 (owner: 10Ahmon Dancy) [19:56:51] (03PS1) 10Ahmon Dancy: jjb: Rename scap-beta-deb to beta-build-scap-deb [integration/config] - 10https://gerrit.wikimedia.org/r/678934 [20:00:11] 10MediaWiki-Releasing, 10MW-1.36-release: Announce 1.36 retrospective "pencils down" and likely timeline to wikitech-l etc. - https://phabricator.wikimedia.org/T279708 (10Jdforrester-WMF) a:03thcipriani [20:01:27] James_F: re: task, sure. :) [20:01:40] i think i was thinking of selenium - https://www.mediawiki.org/wiki/Selenium/Getting_Started/Run_tests_targeting_MediaWiki-Docker [20:01:45] https://phabricator.wikimedia.org/T247312 [20:35:19] brennen: Oh, interesting. [20:48:36] "install a cluster of wikis" -- when docker can do that then I think I can say that mw-vagrant is dead tech ;) [20:55:02] longma: :D [20:55:22] 🥳 [20:56:09] 10Release-Engineering-Team-TODO (2021-04-01 to 2021-06-30 (Q4)), 10Release, 10Train Deployments: 1.37.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T278345 (10Urbanecm) [20:59:34] 10VPS-project-Codesearch, 10Patch-For-Review, 10User-Ladsgroup: Codesearch (Beta) not reachable - https://phabricator.wikimedia.org/T277459 (10Legoktm) Sorry, I had set this up in a very adhoc way originally. Now that it's puppetized it should hopefully not randomly go down/disappear. [21:01:01] 10MediaWiki-Releasing, 10MediaWiki-Installer: Enhance installer to support dependencies between skins and extensions - https://phabricator.wikimedia.org/T196670 (10CCicalese_WMF) Yes, definitely, it should be moved to 1.37 (at a minimum). I'm going to untag it from a specific release, since it is not on any ro... [21:31:07] 10Release-Engineering-Team-TODO (2021-04-01 to 2021-06-30 (Q4)), 10Release, 10Train Deployments: 1.37.0-wmf.1 deployment blockers - https://phabricator.wikimedia.org/T278345 (10Urbanecm) [21:38:16] (03CR) 10Legoktm: [C: 03+1] "The core change has landed now." [tools/scap] - 10https://gerrit.wikimedia.org/r/677616 (https://phabricator.wikimedia.org/T263872) (owner: 10Ahmon Dancy) [22:04:50] The eagle has landed. [22:25:09] James_F: theoretically you could do that sort of integration test with helm test once we have an official mediawiki chart [22:25:30] (if your service runs in k8s) [22:25:48] and if ci runs k8s? [22:26:06] we already support helm test in the pipeline [22:26:26] it deploys to the ci namespace in the staging cluster [22:26:36] or did you mean something else? [22:26:46] nope, that I think :) [22:26:49] :) [22:27:20] I really need to come to one of your office hours, but its at the same time as one of my team meetings every week [22:27:54] those precious east/west working overlap hours... [22:28:35] oh hmm...maybe we could change around the meeting times on occasion. That might help for other people too...or feel free to schedule a 1:1 meeting any time! [22:28:51] haven't been getting much attendance at the office hours :P [22:34:58] I suppose it would be possible to use the mediawiki-dev chart buuut I'm not sure if that's desireable [22:44:48] longma: Yeah, when ins the official m8s chart coming along? :-) [22:46:21] it is here, yet to be merged: https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/670220 I assume it needs a bit more work [22:46:39] Ooh. [22:46:41] * James_F stalks. [22:46:54] :D [23:07:52] 10MediaWiki-Releasing, 10MediaWiki-Installer: Enhance installer to support dependencies between skins and extensions - https://phabricator.wikimedia.org/T196670 (10Jdlrobson) [23:08:23] 10MediaWiki-Releasing, 10MW-1.37-release, 10MinervaNeue (Tracking), 10Readers-Web-Backlog (Tracking): Bundle MinervaNeue skin with MediaWiki - https://phabricator.wikimedia.org/T191743 (10Jdlrobson) [23:08:26] 10MediaWiki-Releasing, 10MW-1.37-release, 10MobileFrontend (Tracking), 10Readers-Web-Backlog (Tracking): Bundle MobileFrontend extension with MediaWiki - https://phabricator.wikimedia.org/T191734 (10Jdlrobson) [23:16:52] 10Continuous-Integration-Infrastructure: Post-merge builds failing with "Command 'mwext-phpunit-coverage' returned non-zero exit status 1" - https://phabricator.wikimedia.org/T280064 (10Daimona) Most probably related to the ongoing work at T234020; specifically, I guess it could be related to T279833. Should be... [23:17:08] 10Continuous-Integration-Infrastructure: Post-merge builds failing with "Command 'mwext-phpunit-coverage' returned non-zero exit status 1" - https://phabricator.wikimedia.org/T280064 (10Daimona) [23:17:11] 10Continuous-Integration-Config, 10Release-Engineering-Team (CI & Testing services), 10Release-Engineering-Team-TODO, 10Patch-For-Review, 10Test-Coverage: Switch mediawiki code coverage from xdebug to pcov - https://phabricator.wikimedia.org/T234020 (10Daimona) [23:42:45] 10Release-Engineering-Team-TODO (Yak Shaving 🐃🪒 ), 10Scap: Automate scap prep - https://phabricator.wikimedia.org/T245184 (10thcipriani) [23:44:58] 10Release-Engineering-Team-TODO (Yak Shaving 🐃🪒 ): Yak: Gerrit user page graph - https://phabricator.wikimedia.org/T280091 (10thcipriani) [23:48:02] 10Release-Engineering-Team-TODO (Yak Shaving 🐃🪒 ): Yak: automate the use of https://releng.team - https://phabricator.wikimedia.org/T280093 (10thcipriani) [23:49:44] 10Release-Engineering-Team-TODO (Yak Shaving 🐃🪒 ): Yak: Investigate HTTPBB during scap - https://phabricator.wikimedia.org/T280094 (10thcipriani) [23:53:05] 10Release-Engineering-Team (Deployment services), 10Release-Engineering-Team-TODO (Yak Shaving 🐃🪒 ), 10Scap, 10serviceops: Define a mediawiki "version" - https://phabricator.wikimedia.org/T218412 (10thcipriani)