Page MenuHomeMiraheze

Run fixStuckGlobalRename.php for R4356th
Closed, ResolvedPublic

Description

It's been nearly an hour and @R4356th's global rename request appears stuck on afterlifewiki. It's not a deleted wiki, so it's not that bug, but could just be related to general sluggishness / backlog on the jobrunners and/or also a lot of attached wikis for the user. Thus, I kindly request that SRE run fixStuckGlobalRename.php for @R4356th (old username R4356th and new username Redmin). Thank you.

Additionally, if you could also link to [[phab:T7736|T7736]] by adding [[phab:T7736|T7736]] to your Tech:Server admin log log entry, that'd be great. Thank you also.

Event Timeline

Dmehus triaged this task as Normal priority.Aug 2 2021, 23:09
Dmehus created this task.
Dmehus moved this task from Backlog to Maintenance Script Run on the MediaWiki board.
Dmehus updated the task description. (Show Details)

An hour is nothing really for a large account. If no one beats me to it, I'll check in the morning.

We seem to having trouble though with load so wouldn't shock me at all.

There's 83 jobs queued just on meta. The rename isn't stuck. The whole queue is.

RhinosF1 changed the task status from Open to Stalled.Aug 2 2021, 23:17

Please re-open if it doesn't start moving when the queue clears. By the constant groaning of icinga, MediaWiki is having issues with load.

We have nearly 10k jobs. running them but it's not gonna be fast.

We have nearly 10k jobs. running them but it's not gonna be fast.

@RhinosF1 I would've still preferred you run fixStuckGlobalRename.php first, as it would get rid of the global rename jobs in one, quick swoop, then ran jobs on all wikis. This way, there's no guarantee the global rename will be fixed by tomorrow, and the user will be inconvenienced by not being able to use their user account. That's what @Reception123 has always done for me, and plus, since @John fixed the way the jobs are handled in terms of a lot of "abandoned" global user page update jobs on loginwiki, global renames have typically been completed within minutes—when with a lot of attached wikis and/or global edits. The fact the rename appears stuck on a small wiki, afterlifewiki, suggests something more is at play here.

We have nearly 10k jobs. running them but it's not gonna be fast.

@RhinosF1 I would've still preferred you run fixStuckGlobalRename.php first, as it would get rid of the global rename jobs in one, quick swoop, then ran jobs on all wikis. This way, there's no guarantee the global rename will be fixed by tomorrow, and the user will be inconvenienced by not being able to use their user account.

1 user that's asleep doesn't take priority over the whole site lagging. The jobs script will not take that long.

We have nearly 10k jobs. running them but it's not gonna be fast.

@RhinosF1 I would've still preferred you run fixStuckGlobalRename.php first, as it would get rid of the global rename jobs in one, quick swoop, then ran jobs on all wikis. This way, there's no guarantee the global rename will be fixed by tomorrow, and the user will be inconvenienced by not being able to use their user account.

1 user that's asleep doesn't take priority over the whole site lagging. The jobs script will not take that long.

But you said to ping you tomorrow morning (i.e., 8-12 hours) if it's not done. I really shouldn't have mentioned the outstanding jobs as the likely cause and you've grabbed onto that as a jobs issue, but I suspect, rather strongly, something else is at play here. It's no sweat, and less taxing on the server, to run fixStuckGlobalRename.php, than to run jobs on all wikis.

Dmehus changed the task status from Invalid to Declined.Aug 2 2021, 23:36

yeah then I decided to stay awake. We're moving fairly fast through the queue. The issues at play are the entire capacity of MediaWiki. It's been tipping for weeks. Between GlobalUserPages and a bunch of Renames for RemovePII testing, it just seems grumpy.

yeah then I decided to stay awake. We're moving fairly fast through the queue. The issues at play are the entire capacity of MediaWiki. It's been tipping for weeks. Between GlobalUserPages and a bunch of Renames for RemovePII testing, it just seems grumpy.

Okay. How about I check again in 3-4 hours and if it's still stuck on the same wiki, I'll reopen the task?

If we're still this bad in 3-4 hours, they'll be an Unbreak now incident.

We're seeing an actual failure now so the wiki does need a repair.

We're seeing an actual failure now so the wiki does need a repair.

Yeah, just came here to link to that. Thanks!

If we're still this bad in 3-4 hours, they'll be an Unbreak now incident.

heh, true. Fair enough. ;)

RhinosF1 raised the priority of this task from Normal to High.Aug 2 2021, 23:58

A wiki is giving the NameTableAccess nonsense again, normal steps not working and the script fails.

Void claimed this task.
Void subscribed.

Ran /srv/mediawiki/w/maintenance/populateContentTables.php --wiki=afterlifewiki and then fixStuckGlobalRename.php and it unstuck normally.