[04:42:43] <Tulsi>	 o/ AlexZ, a friendly reminder to deploy Wikimedia cloak. Its been 34 days since i request for it. :)
[06:01:31] <PleaseStand>	 TimStarling: Is there a reason the populateEditCount.php script used for board elections doesn't do waitForReplication()?
[06:01:44] <PleaseStand>	 legoktm said I should ask you because you ran it last year
[06:10:00] <TimStarling>	 no
[06:12:13] <PleaseStand>	 I mean, preparing for elections always happens at the last minute, will waitForReplication() slow it down too much?
[06:15:19] <PleaseStand>	 TimStarling: ^
[06:16:31] <TimStarling>	 just do it when $numUsers % 100 == 0
[06:17:14] <TimStarling>	 if it actually caused replication lag I would have added it, but there is no particular reason why it's not there already
[06:18:10] <PleaseStand>	 So you think 500 is OK? https://gerrit.wikimedia.org/r/c/mediawiki/extensions/SecurePoll/+/814346
[06:19:47] <TimStarling>	 yes, gave +2
[06:20:44] <PleaseStand>	 Thank you for your input TimStarling
[06:23:16] <TimStarling>	 the script is maybe older than you think, I updated the select queries recently but the insert and general structure goes back to 2009
[06:26:15] <TimStarling>	 https://static-codereview.wikimedia.org/MediaWiki/53859.html
[06:47:58] <PleaseStand>	 foks, TimStarling: I added the patch to the deployment calendar
[06:48:27] <PleaseStand>	 To avoid going through code deployment every election year, do you think it would make sense for me to create a generic version of the script for next time?
[06:54:46] <foks>	 PleaseStand, probably :D 
[06:54:58] <foks>	 Didn't it go out on Monday? Or am I misreading?
[06:55:44] <foks>	 Oh, Z_abe's patch needed merged, that's right
[06:58:13] <PleaseStand>	 foks: And assuming that you want to run the script on all wikis, I think it needs to exist in all active deployment branches
[06:58:25] <PleaseStand>	 https://noc.wikimedia.org/conf/ Currently active MediaWiki versions: php-1.39.0-wmf.19, php-1.39.0-wmf.21
[06:59:42] <foks>	 ah I see, yeah
[06:59:51] * foks is pretty new to this!
[16:30:12] <PotsdamLamb>	 is the toolforge login not working? I keep getting "upstream request timeout"
[16:36:31] <taavi>	 PotsdamLamb: login where exactly?
[16:37:12] <PotsdamLamb>	 https://toolsadmin.wikimedia.org/register/
[16:37:39] <PotsdamLamb>	 https://toolsadmin.wikimedia.org/ won't work either
[16:37:56] <taavi>	 oh, to toolsadmin and not via ssh
[16:38:07] <taavi>	 bd808: ^ maybe related to your striker work recently?
[16:38:09] <PotsdamLamb>	 correct
[16:38:23] <PotsdamLamb>	 striker?
[16:38:44] <taavi>	 'striker' is what we call toolsadmin.wikimedia.org internally
[16:39:07] <bd808>	 hmm... it could be broken for sure. I changed the hosting yesterday.
[16:39:30] <bd808>	 "upstream request timeout" sounds like a problem between the CDN edge and the app
[16:39:35] <taavi>	 I don't seem to be able to login either
[16:39:41] <PotsdamLamb>	 hhhmmmmmmm
[16:40:17] <PotsdamLamb>	 sounds like my old engineers lets make a change, not test it and see if anyone has an issue. I am heading out for a 6 hour lunch lol
[16:40:18] <taavi>	 yeah, but I suspect some part of the cdn->envoy->uwsgi-in-docker has a lower timeout than either uwsgi or some external resource it tries to connect to
[16:40:19] <bd808>	 I'm in a meeting, but I can dig into this in about 20 minutes
[16:40:28] <PotsdamLamb>	 ok thanks @bd808
[17:47:38] <bd808>	 PotsdamLamb, taavi: things seem to be better for now following a docker container restart on both backend servers. I will spend some more time staring at logs after my lunch.
[17:53:33] <PotsdamLamb>	 ok it works :)
[18:39:32] <legoktm>	 PleaseStand: ty for following up! I agree that having a single script is good, though if the board changes eligibility requirements, it would need a code change. (I think I prefer hardcoding the requirements in the script itself rather than having them be CLI parameters just so it's easier for everyone to verify the correct values were used)
[18:41:54] <legoktm>	 PleaseStand: also, maybe we should have the script be a no-op on loginwiki, last time Tim had to kill the script manually since it has a ton of users but will never have any eligible users.