[00:30:10] Hello everyone, does anyone know if there is anyway to rename the "Discussion" tab to something else? for e.g. instead of "Discussion", I want it to show as "Collaboration" [00:30:36] By "Discussion" tab I mean the one that shows on top of every page [00:37:16] Guest50: edit "MediaWiki:Talk" on your wiki [00:38:00] see https://www.mediawiki.org/wiki/Help:System_message#Overriding_messages_on-wiki [00:46:34] Thank you! that is simple! [01:08:04] :) [04:14:26] Is there a way to find out which pages of the wiki are taking the longest to parse and serve (create the HTML from a cold cache)? [10:41:32] taylan: I don't think so, no. At least not without some manual patching of the code. [13:25:04] Krinkle: hi, I'm wondering if you know why keys such as "global:jobqueue:sonicpediawiki:htmlCacheUpdate:rootjob:2c5890b064405e3af7397907047e570702e9b556" (rootjob) just stay? Nothing is done and it's not removed and stays permanently. [13:57:00] paladox: are you using redis for the job queue? You may need to set up this https://www.mediawiki.org/wiki/Redis#Automatic_handling_of_job_recycling_and_abandons [14:16:25] Yes [14:17:00] oh that's interesting [14:18:21] Vulpix: so I need to create a jobrunner-cleanup service? Which is similar to jobrunner apart from different config. I suppose I could run that on the jobchron seeing as it requires no runJobs.php [14:24:51] oh, jobchron needs to be using that config not jobrunners [14:25:05] so I don't need to create two jobchron service files using different configs [14:26:55] paladox: I use 2 services, one for runJobs.php and a second one for this redisJobChronService [14:27:47] If you have multiple wikis using the same redis instance, you only need to run one instance of redisJobChronService, pointing to the redis instance that stores the jobs [14:34:04] Vulpix: ah thanks! We have two services. We run jobrunner on multiple servers then we have jobchron which we run on the server that has redis . jobchron was using the same config as jobrunner. But I think with the link you gave me, I can just use that one. [14:34:28] https://github.com/miraheze/puppet/blob/paladox-patch-12/modules/mediawiki/templates/initscripts/jobchron.systemd.erb [14:39:41] although [14:39:43] > jobchron121 jobchron[3155800]: 2023-11-07T14:28:39+0000 NOTICE: Updated the state of 1 job(s) (recycle/undelay/abandon). [14:39:54] seems I kinda have it already? [14:48:13] Yes, that's the usual redisJobChronService output [15:03:39] seems it's not removing the rootjobs tho [15:11:26] Vulpix: does rootjobs build up for you? [15:11:44] You're right. I have 28000+ *rootjob* redis keys [15:16:43] I guess there could be a bug? [15:18:54] Indeed. I remember when I discovered this bug https://gerrit.wikimedia.org/r/c/mediawiki/core/+/879656 and since WMF doesn't use redis anymore, it's more likely that something just got broken somehow [15:19:50] There are some complex logic with lua going on there, which makes it difficult to debug [15:24:34] paladox: Is there a phabricator task for this? I may be interested in taking a look when I have time, if I discover something (I'll be busy this week) [15:24:46] No rn. I can open one tho [15:26:20] https://phabricator.wikimedia.org/T350696 [15:33:04] thanks! [20:52:53] I guess some how in here https://github.com/wikimedia/mediawiki/blob/master/includes/jobqueue/JobQueueRedis.php#L390 it should be deleting the rootjob?