[07:40:31] 10serviceops, 10OpenRefine, 10Operations, 10Traffic, 10Core Platform Team Workboards (Clinic Duty Team): Clients failing API login due to dependence on "Set-Cookie" header name casing - https://phabricator.wikimedia.org/T249680 (10MoritzMuehlenhoff) p:05Triage→03High [07:53:34] 10serviceops, 10OpenRefine, 10Operations, 10Traffic, 10Core Platform Team Workboards (Clinic Duty Team): Clients failing API login due to dependence on "Set-Cookie" header name casing - https://phabricator.wikimedia.org/T249680 (10Joe) As far as I can see, apache (which sits beyond envoy) emits all cooki... [08:06:33] 10serviceops, 10OpenRefine, 10Operations, 10Traffic, 10Core Platform Team Workboards (Clinic Duty Team): Clients failing API login due to dependence on "Set-Cookie" header name casing - https://phabricator.wikimedia.org/T249680 (10Pintoch) Switching on headers capitalization would be absolutely fantastic... [09:54:40] 10serviceops, 10MediaWiki-API, 10Operations, 10Core Platform Team Workboards (External Code Reviews), 10MW-1.35-notes (1.35.0-wmf.27; 2020-04-07): CORS errors on commons on debug servers - https://phabricator.wikimedia.org/T249107 (10Tgr) 05Open→03Resolved a:03Tgr The error can't be reproduced on C... [10:13:30] <_joe_> hnowlan: I added a couple comments to https://gerrit.wikimedia.org/r/#/c/operations/deployment-charts/+/587298/ [10:13:47] <_joe_> I'm happy to discuss the substance if you have doubts [11:47:53] _joe_: thanks! that CR has been merged, it was accidentally reopened I think. But the points are still valid - I'll test and add them to a new change [13:17:28] 10serviceops, 10Growth-Team, 10MediaWiki-Configuration, 10Parsoid, and 4 others: pywikibot encounters an internal API error with Flow on testwiki (but not other wikis) - https://phabricator.wikimedia.org/T249705 (10Joe) [13:26:54] 10serviceops, 10Growth-Team, 10MediaWiki-Configuration, 10Parsoid, and 5 others: pywikibot encounters an internal API error with Flow on testwiki (but not other wikis) - https://phabricator.wikimedia.org/T249705 (10kostajh) > I just found your errors in the logs, and this thickens the cloud of mistery arou... [14:18:20] _joe_: seems that loading the file out of /etc won't work, helm limits access to the chart. Unless I misunderstood your idea and we should copy the cert to the chart itself as a file and load it that way as opposed to out of values.yaml [14:19:07] <_joe_> hnowlan: well or we can just have puppet copy the file over [14:19:33] <_joe_> hnowlan: I'm fixing a patch, then I can assist with that [14:19:46] puppet would copy it into the helmfile directory? [14:19:48] ah, cool [14:20:16] <_joe_> hnowlan: or it can read it and copy it in a proper place, but yes it's annoying [14:20:44] <_joe_> hnowlan: so either it's in the chart, or it can't be read with Files.Get? [14:20:55] <_joe_> meh [14:21:00] <_joe_> it also kinda-makes sense [14:21:09] <_joe_> so, uhm. [14:21:09] yeah, tiller doesn't know what's on the filesystem on install [14:22:06] <_joe_> so, gimme 10 minutes and I can try to help you get unblocked [14:22:20] cool, thanks! [14:22:29] <_joe_> the solution will be injecting it in a values.yaml-like file [14:22:33] <_joe_> from puppet [14:24:51] <_joe_> and sorry for sending you down the wrong implementation path [14:25:11] <_joe_> this is something we'd solve if we had a CI system to build the charts [14:26:40] <_joe_> hnowlan: so let's look at puppet class profile::kubernetes::deployment_server::helmfile [14:27:45] <_joe_> look at line 83 [14:28:06] <_joe_> there is a template explicitly created to do what we need - pass values from puppet to the charts [14:28:44] <_joe_> we can probably add the puppet CA for all charts :) [14:28:58] oh, nice! [14:29:23] or at least better - it still feels kinda ugly but it's better than doing it 1000 times [14:30:00] <_joe_> it's also better that it will change whenever the puppet ca will change [14:30:32] yeah, that'll be good [14:30:40] <_joe_> I don't remember if the CA is available in puppet as a variable [14:30:48] <_joe_> or if you will have to read the file somehow [14:31:23] I'll have a look [15:54:40] 10serviceops, 10Operations, 10Kubernetes: New Deployment charts should allow exposing services via TLS - https://phabricator.wikimedia.org/T236008 (10Joe) 05Open→03Resolved [15:54:43] 10serviceops, 10Operations, 10Kubernetes, 10Patch-For-Review: Add TLS termination to services running on kubernetes - https://phabricator.wikimedia.org/T235411 (10Joe) [15:55:06] 10serviceops, 10Operations, 10Release-Engineering-Team: Hundreds of tags for `wikimedia/mediawiki-core` image - https://phabricator.wikimedia.org/T242775 (10Joe) p:05High→03Low [16:08:50] James_ F says 'probably' (the Tuesday preceding the 5 day stretch) will be a no-deploy day but I am waiting for greg -g to confirm. [16:35:12] 10serviceops, 10OpenRefine, 10Operations, 10Traffic, and 2 others: Clients failing API login due to dependence on "Set-Cookie" header name casing - https://phabricator.wikimedia.org/T249680 (10Joe) 05Open→03Resolved a:03Joe From my tests, now we get all cookies correctly set: ` $ curl --http1.1 -sIL... [16:36:44] 10serviceops, 10OpenRefine, 10Operations, 10Traffic, and 2 others: Clients failing API login due to dependence on "Set-Cookie" header name casing - https://phabricator.wikimedia.org/T249680 (10Joe) @Pintoch the behaviour should be restored now. Can you confirm old versions of OpenRefine work correctly now? [16:55:15] greg -g says: apergos: tl;dr: no train that week and no deploys on those days off. swats open for the non-off days though. cc thci priani (right?) [16:55:35] so now waiting for thcipri ani to agree or not [16:56:44] 10serviceops, 10OpenRefine, 10Operations, 10Traffic, and 2 others: Clients failing API login due to dependence on "Set-Cookie" header name casing - https://phabricator.wikimedia.org/T249680 (10Pintoch) Thanks a million, this is very kind of you! I can confirm this works, edits are coming through again fro... [17:05:09] 10serviceops, 10MediaWiki-Parser, 10Operations, 10Core Platform Team Workboards (Clinic Duty Team), and 2 others: API action=parse should be poolcounter-limited if a re-parse is necessary - https://phabricator.wikimedia.org/T243803 (10Anomie) Next step is to determine the limits and implement them for Wiki... [18:17:57] 10serviceops, 10Growth-Team, 10MediaWiki-Configuration, 10Parsoid, and 5 others: pywikibot encounters an internal API error with Flow on testwiki (but not other wikis) - https://phabricator.wikimedia.org/T249705 (10Dvorapa) Just for the record, here is the error after @ssastry (? I guess) edited that parti... [18:20:17] 10serviceops, 10MediaWiki-Page-derived-data, 10Performance-Team: Post-send DeferredUpdates unreliable due to fpm timeout before MW (missing edits in Watchlist and Recentchanges) - https://phabricator.wikimedia.org/T248564 (10Krinkle) [19:59:46] (09:31:23 μμ) greg -g: tl;dr: no train that week and no deploys on those days off. swats open for the non-off days though. -- verified by thcipri ani [20:00:18] new calendar should be up Soon™ [23:52:54] 10serviceops, 10Growth-Team, 10MediaWiki-Configuration, 10Parsoid, and 5 others: pywikibot encounters an internal API error with Flow on testwiki (but not other wikis) - https://phabricator.wikimedia.org/T249705 (10Dvorapa) Occasionally also fails on mediawiki.org (https://travis-ci.org/github/wikimedia/py...