[13:30:50] !log admin creating indexes and views for banwiki T234770 [13:30:53] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Admin/SAL [13:30:53] T234770: Prepare and check storage layer for banwiki - https://phabricator.wikimedia.org/T234770 [17:10:54] !log tools restart tools-worker-1035 because it is no longer responding [17:10:57] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [18:25:18] zhuyifei1999_ find an evening, let me know when you are free, and let [18:25:32] us have a nice buster migration party [18:25:39] what is your timezone? [18:25:47] UTC +3 [18:25:55] moving to +2 in a few weeks [18:26:20] UTC - 5 here hmm [18:26:48] just half a world apart, not too complicated ;) [18:27:59] utc -5 is mountain time? [18:28:46] no, central time [18:29:42] oh. Need to plan a trip to the states. Many open promises I need to execute on [18:29:42] how about 17 UTC / later on Oct 26 / 27? [18:29:47] lol [18:30:10] Better towards the 26/27, I am on vacation until the 23rd [18:30:11] I'm unfortunately busy this weekend [18:30:15] ok [18:31:21] I will poke you here on those dates in my evenings, let's see what works out. Thank you being awesome! [18:31:31] *for [18:31:53] k :) [18:32:00] * matanya wonders away to fight some nudgy vandals [18:34:31] so, we have a pretty simple VPS running MediaWiki at https://thegoodplace.wmflabs.org/ for some reason, trying to access https://thegoodplace.wmflabs.org/wiki/Special:UserLogin in Chrome results in a net::ERR_HTTP2_PROTOCOL_ERROR 200 [18:34:53] this doesn't happen in Firefox [18:36:53] I the request seems to make it to Apache (as HTTP/1.1): [15/Oct/2019:18:35:03 +0000] "GET /wiki/Special:UserLogin HTTP/1.1" 200 4414 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/77.0.3865.90 Safari/537.36" [18:38:16] what's really weird is that static (cached?) pages seem to be fine, like: https://thegoodplace.wmflabs.org/wiki/Wednesday [18:40:12] davidwbarratt: do you think this is likely to be a Cloud VPS front proxy problem or a MediaWiki problem? [18:40:55] I think the former since I haven't seen this in production or locally [18:41:11] the pages that seem to fail are the ones that set a session cookie.... [18:41:34] I can create a ticket if you would like [18:42:06] that would be a good start [18:42:53] I guess it could be a problem with MediaWiki, like a permission error when attempting to save the session or something... [18:43:00] but it works in Firefox [18:43:05] so that kinds spoils that theory [18:44:02] davidwbarratt: is thegoodplace running as a mediawiki-vagrant deployment or by some other means? [18:44:32] it's running the base docker image: https://hub.docker.com/_/mediawiki/ [18:45:55] not super helpful SO question that seems related -- https://stackoverflow.com/questions/58215104/whats-the-neterr-http2-protocol-error-about [18:47:07] bd808 haha, yeah I saw that one, but was like... really... that's the accepted answer?? haha [18:50:25] I fear that someone is going to have to do some digging with Chrome netlog to figure out what it doesn't like -- https://www.chromium.org/developers/design-documents/network-stack/netlog [18:50:53] ok, I tested Safari and that browser is fine too, so it looks like it's a Chrome only issue [19:24:53] bd808 could I try deleting the web proxy and recreating it? or would that be a bad idea? [19:25:48] davidwbarratt: its not a bad idea technically. All that does is remove some data from redis and then add some data to redis, so I'm not highly confident it will change anything. [19:26:03] hmm [19:27:07] yeah no change [19:53:10] bd808 here ya go! https://phabricator.wikimedia.org/T235554