[00:18:53] <legoktm>	 can I build pipeline images using a branch for local testing?
[00:19:43] <legoktm>	 for example, if I could get https://gerrit.wikimedia.org/r/c/mediawiki/libs/Shellbox/+/737995/ (switch to PHP 7.4) merged in some fashion, but without blocking all future patches if I'm not ready to deploy it yet
[00:20:37] <legoktm>	 I suppose I could just publish php74 variants for all of them
[00:20:57] <legoktm>	 and just live with duplication until we're ready to drop 72 fully
[00:35:52] <wikibugs>	 10Release-Engineering-Team (Doing), 10Scap, 10MediaWiki-Release-Tools, 10Patch-For-Review: Output of scap lint.py isn't helpful if it fails the php -l check - https://phabricator.wikimedia.org/T272760 (10Reedy) >>! In T272760#7497476, @dancy wrote: > I ran into this today while experimenting, so I fixed it...
[06:58:38] <majavah>	 could a phan admin delete https://phabricator.wikimedia.org/T162836#7497884 please
[07:09:04] <wikibugs>	 10Phabricator (Upstream), 10Developer Productivity, 10Upstream: Disable parent task metadata by default for new sub tasks - https://phabricator.wikimedia.org/T239378 (10Aklapper) Implementing this would require some code changes, so in my understanding this would fall into the #release-engineering-team realm.
[08:28:32] <wikibugs>	 (03PS1) 10Kosta Harlan: Browser tests: run npm install in parallel [integration/quibble] - 10https://gerrit.wikimedia.org/r/738061 (https://phabricator.wikimedia.org/T226869)
[08:32:00] <wikibugs>	 (03PS1) 10Kosta Harlan: [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066
[08:41:45] <wikibugs>	 (03CR) 10jerkins-bot: [V: 04-1] [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066 (owner: 10Kosta Harlan)
[09:03:20] <wikibugs>	 (03PS2) 10Kosta Harlan: [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066
[09:08:18] <wikibugs>	 (03CR) 10jerkins-bot: [V: 04-1] [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066 (owner: 10Kosta Harlan)
[09:19:16] <wikibugs>	 (03PS3) 10Kosta Harlan: [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066
[09:28:05] <wikibugs>	 (03PS4) 10Kosta Harlan: [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066
[09:48:19] <wikibugs>	 (03PS1) 10Kosta Harlan: [DNM] Selenium test run [integration/quibble] - 10https://gerrit.wikimedia.org/r/738184
[11:16:12] <wikibugs>	 (03CR) 10Kosta Harlan: "The comparison of build times:" [integration/quibble] - 10https://gerrit.wikimedia.org/r/738061 (https://phabricator.wikimedia.org/T226869) (owner: 10Kosta Harlan)
[11:21:51] <_joe_>	 I might break beta's mediawiki briefly. Hopefully just seconds, worst case scenario a few minutes
[11:22:31] <_joe_>	 actually, let me spin up a new VM, so I can also test a clean-slate install first
[11:27:45] <wikibugs>	 (03PS2) 10Kosta Harlan: BrowserTests: Option to run npm install in parallel [integration/quibble] - 10https://gerrit.wikimedia.org/r/738061 (https://phabricator.wikimedia.org/T226869)
[11:30:29] <wikibugs>	 (03PS3) 10Kosta Harlan: BrowserTests: Option to run npm install in parallel [integration/quibble] - 10https://gerrit.wikimedia.org/r/738061 (https://phabricator.wikimedia.org/T226869)
[11:36:26] <wikibugs>	 (03PS4) 10Kosta Harlan: BrowserTests: Option to run npm install in parallel [integration/quibble] - 10https://gerrit.wikimedia.org/r/738061 (https://phabricator.wikimedia.org/T226869)
[11:41:35] <wikibugs>	 10Release-Engineering-Team (Done by Wed 24 Nov 🔥), 10Release, 10Train Deployments: 1.38.0-wmf.9 deployment blockers - https://phabricator.wikimedia.org/T293950 (10Lucas_Werkmeister_WMDE)
[11:59:17] <wikibugs>	 (03Abandoned) 10Kosta Harlan: [DNM] Selenium test run [integration/quibble] - 10https://gerrit.wikimedia.org/r/738184 (owner: 10Kosta Harlan)
[11:59:23] <wikibugs>	 (03Abandoned) 10Kosta Harlan: [DNM] Add more dependencies to full run to check build time [integration/quibble] - 10https://gerrit.wikimedia.org/r/738066 (owner: 10Kosta Harlan)
[13:05:17] <wikibugs>	 10Release-Engineering-Team (Deployment Training Requests): Deployment training request for scherukuwada - https://phabricator.wikimedia.org/T293087 (10SCherukuwada) Yes, I did attend one session where I got walked through a deployment. It was very informative. I still have to actually deploy once myself though.
[14:54:12] <wikibugs>	 10Release-Engineering-Team (Deployment Training Requests): Deployment training request for scherukuwada - https://phabricator.wikimedia.org/T293087 (10Aklapper) If the deployment training took place, then the status of this ticket should be set to `resolved`, I assume.
[14:55:58] <wikibugs>	 10Release-Engineering-Team (Doing), 10Scap, 10MediaWiki-Release-Tools, 10Patch-For-Review: Output of scap lint.py isn't helpful if it fails the php -l check - https://phabricator.wikimedia.org/T272760 (10dancy) >>! In T272760#7497584, @Reedy wrote: >>>! In T272760#7497476, @dancy wrote: >> I ran into this...
[14:57:48] <wikibugs>	 10Release-Engineering-Team (Doing), 10Scap, 10MediaWiki-Release-Tools, 10Patch-For-Review: Output of scap lint.py isn't helpful if it fails the php -l check - https://phabricator.wikimedia.org/T272760 (10Reedy) >>! In T272760#7498901, @dancy wrote: >>>! In T272760#7497584, @Reedy wrote: >>>>! In T272760#74...
[15:05:16] <wikibugs>	 (03PS1) 10Kosta Harlan: Adjust cache settings for more performant browser tests [integration/quibble] - 10https://gerrit.wikimedia.org/r/738252
[15:14:05] <wikibugs>	 (03PS1) 10Kosta Harlan: [DNM] Checking cache changes with more dependencies [integration/quibble] - 10https://gerrit.wikimedia.org/r/738253
[15:19:06] <wikibugs>	 10Beta-Cluster-Infrastructure, 10Patch-For-Review: deployment-deploy* failing puppet due to kafka not having ipv6 - https://phabricator.wikimedia.org/T281986 (10Majavah) 05Open→03Resolved a:03Majavah
[15:21:57] <wikibugs>	 10Beta-Cluster-Infrastructure, 10Release-Engineering-Team (Radar): Migrate deployment-prep away from Debian Stretch to Buster/Bullseye - https://phabricator.wikimedia.org/T278641 (10Majavah)
[15:53:58] <icinga-wm>	 PROBLEM - SSH on contint1001.mgmt is CRITICAL: CRITICAL - Socket timeout after 10 seconds https://wikitech.wikimedia.org/wiki/Dc-operations/Hardware_Troubleshooting_Runbook
[16:55:00] <icinga-wm>	 RECOVERY - SSH on contint1001.mgmt is OK: SSH OK - OpenSSH_6.6 (protocol 2.0) https://wikitech.wikimedia.org/wiki/Dc-operations/Hardware_Troubleshooting_Runbook
[17:39:11] <Krinkle>	 kostajh: I've thus far shied away from (trying to) further optimise MW in CI because the fresnel jobs suggest that we're pretty much as fast as can do (hundred serial page views and dozens chromium restarts all under a minute in total. Same MW/php/Quibble/MySQL settings.
[17:40:19] <Krinkle>	 After the use of MySQL on ram/tmpfs and enabling of general caching, it's gotten pretty fast with a few ms per http request.
[17:41:00] <Krinkle>	 So my money is on there being something in the webdriver and/or wdio layers and/or the way we use them that makes it so absurdly slow
[17:41:22] <Krinkle>	 Having said that, we can give it a go of course.
[17:43:13] <Krinkle>	 I wonder what it would take to capture an excimer debug flamegraph from a quibble served request
[17:43:34] <Krinkle>	 (iff the time is spent in http/php)
[17:53:36] <kostajh>	 Krinkle: yeah, I'm not sure adjusting the cache settings will make that much of a difference here, just an experiment. It does look like parallel npm install helps, though https://gerrit.wikimedia.org/r/c/integration/quibble/+/738061/
[17:56:22] <wikibugs>	 (03CR) 10Krinkle: Adjust cache settings for more performant browser tests (033 comments) [integration/quibble] - 10https://gerrit.wikimedia.org/r/738252 (owner: 10Kosta Harlan)
[17:56:41] <wikibugs>	 (03CR) 10Kosta Harlan: [C: 04-2] "Meh, Ifabd9f1cb9b0813ccff5dd6d534073feec7c209b (more extensions, without these settings) is 18m 13s, while the same run with these setting" [integration/quibble] - 10https://gerrit.wikimedia.org/r/738252 (owner: 10Kosta Harlan)
[17:57:54] <wikibugs>	 (03PS2) 10Kosta Harlan: Adjust cache settings for more performant browser tests [integration/quibble] - 10https://gerrit.wikimedia.org/r/738252
[17:58:00] <wikibugs>	 (03PS2) 10Kosta Harlan: [DNM] Checking cache changes with more dependencies [integration/quibble] - 10https://gerrit.wikimedia.org/r/738253
[17:58:03] <wikibugs>	 (03CR) 10Kosta Harlan: [C: 04-2] Adjust cache settings for more performant browser tests (033 comments) [integration/quibble] - 10https://gerrit.wikimedia.org/r/738252 (owner: 10Kosta Harlan)
[20:24:17] <wikibugs>	 (03PS1) 10Kosta Harlan: dockerfiles: Use opcache optimizations with built-in PHP server [integration/config] - 10https://gerrit.wikimedia.org/r/738281
[20:31:05] <wikibugs>	 (03PS1) 10Kosta Harlan: jjb: Switch integration-quibble-fullrun to 1.2.0-s2 [integration/config] - 10https://gerrit.wikimedia.org/r/738282
[20:38:50] <wikibugs>	 10Phabricator, 10Developer Productivity: Disable parent task metadata by default for new sub tasks - https://phabricator.wikimedia.org/T239378 (10Aklapper)
[21:51:15] <wikibugs>	 (03CR) 10Krinkle: dockerfiles: Use opcache optimizations with built-in PHP server (032 comments) [integration/config] - 10https://gerrit.wikimedia.org/r/738281 (owner: 10Kosta Harlan)
[21:54:58] <wikibugs>	 (03CR) 10Krinkle: dockerfiles: Use opcache optimizations with built-in PHP server (031 comment) [integration/config] - 10https://gerrit.wikimedia.org/r/738281 (owner: 10Kosta Harlan)
[21:57:31] <wikibugs>	 10Phabricator, 10Developer Productivity: Disable parent task metadata by default for new sub tasks - https://phabricator.wikimedia.org/T239378 (10Krinkle) +1 for disabling entirely. In my experience, the sub tasks generally need fewer or different tags. And the awareness and understanding of the (unwanted) tag...