Page MenuHomeMiraheze

[ACCESS REQUEST] Expanded access for Universal Omega
Closed, ResolvedPublic

Description

Shell name: universalomega
Requested access: cache-admins

Rationale for access: To allow for depooling mw servers for reboot (when upgrading, etc) since I have root on mw*, this kinda goes hand-in-hand to allow me to depool servers (using backend.set_health <backend_name> sick (or auto to repool) in varnishadm). The reason I'm creating this task now, is I did reboots on mw* today after upgrading, as I didn't consider that it would potentially cause issues user-side, but it did for about 10 seconds while the server was rebooting (I did do each server separately to try and prevent that, but it still caused at least one user to notice the 503, reported on IRC). It could also be useful for someone else to have the ability to depool/repool servers in the event it is necessary. For example, it may be necessary during MediaWiki upgrades, or if there's a need for debugging something, etc... so it isn't necessary to always rely solely on @Reception123 to do it, as they are the only one in MediaWiki (SRE) with access to right now.

P.S. If this is declined, then it can be closed as such with no arguments from me. This request is just to give me access that would make some things easier, but if it is deemed unnecessary then that's fine also.

Event Timeline

Noting that cache-admins hasn't been used in years, so not sure if it's in a usable state.

Noting that cache-admins hasn't been used in years, so not sure if it's in a usable state.

According to puppet it should be.

Having discussed with John, I've decided to approve this request. @Universal_Omega has done maintenance on MediaWiki servers multiple times and has worked on improving performance and is an active member of the MW (SRE) team. The rationale provided seems reasonable and I think him being able to repool without needing to ask me or a member of the Infrastructure team will be beneficial.