Page MenuHomeMiraheze

Create Wiki Not Working Properly
Closed, ResolvedPublic

Description

11/27 it seems several requesters were not given bureaucrat rights to their newly created wikis. I'm aware of at least three and I suspect there may be several more. I put bureaucrat rights requests on SN for two.

11/28 I created a wiki from https://meta.miraheze.org/wiki/Special:RequestWikiQueue/38074. It generated the wiki in the database but did not fill in the main page boiler plate. It also did not even note the requester as having an account on the wiki, so also no bureaucrat rights.

Event Timeline

Redmin triaged this task as Unbreak Now! priority.Nov 28 2023, 14:35
Redmin added a subscriber: Chicken4War.

If this cannot be fixed soon enough, it would be helpful if wiki creations are disabled to prevent anyone from accidentally creating new wikis (@Jph2 and I have already stopped approving new requests but if someone returns after a break, they may be unaware of this issue).

Weather or not we disable wiki creations, we should put a notice on meta saying that wiki creation is currently not available, & hopefully that will slow things down enough to assign the rights to the users who have already gotten their wikis accepted & not been given the rights. It will also make it so that when it starts working right again Miraheze will not fall behind on the requests.

Void lowered the priority of this task from Unbreak Now! to Normal.Nov 30 2023, 20:34
Void subscribed.

Doesn't seem to be a persistent problem, though it does seem like a few accounts still need permissions. Wondering if there might be something we can do to increase resilience against similar failures in the future.

Also wondering if we have a script or something lying around from the last time this kind of thing happened to make fixing easier. Maybe it could be a good idea to have a script that counts the number of wikis without a bureaucrat (or member of another group with managewiki permissions) and add it to monitoring.

@Void How about a bot with (userrights-interwiki) that periodically patrols the wiki creation log and checks if bureaucrat + sysop was granted, and grants them if not? It would hopefully deal with this before the user realizes that permissions are missing if CreateWiki somehow doesn't grant them, and would work automatically, no need for a Steward to fix this.

@Void How about a bot with (userrights-interwiki) that periodically patrols the wiki creation log and checks if bureaucrat + sysop was granted, and grants them if not? It would hopefully deal with this before the user realizes that permissions are missing if CreateWiki somehow doesn't grant them, and would work automatically, no need for a Steward to fix this.

I would rather resilience be built into Manage/CreateWiki.

The issue of CreateWiki not finishing has returned. Presently, it is not putting the Main Page boilerplate information on the new wiki. It is also failing to assign the requestors as users to their wikis. As a result, it is also not giving bureaucrat rights to requestors.

Hi,

I would like to start by apologizing on the project's behalf that your task was not resolved within a reasonable time. The community has now decided via RfC to merge with WikiTide and former volunteers will be returning and new ones will likely join. A wait such as this one is unlikely to occur again.

Since your task has been opened for a long time, please let us know if (1) you still need it to be done or (2) if it's a bug: if it's still occurring.

There haven't been any issues since a fix was applied mid-December. I think this task can be closed.

This has reared its ugly head again today. I think most wikis created today didn't give permissions to the requestors and at least one didn't even attach the requestor to the new wiki. Some new wikis also didn't get their main page boilerplate.

Unknown Object (User) added a comment.Jan 7 2024, 18:34

That sounds like a Redis issue to me. The only cause I can think of is that Redis ran out of memory or something so not much we can do for now until we migrate to new cloud servers and can allocate mwtask* with more resources.

Closing as resolved as there have been no reported issues since then to my knowledge and given the time that has passed it would be difficult to get logs now to investigate. If this happens again please reopen