Wikimedia Platform Engineering/MediaWiki Core Team/checkin-20130930
Who Bryan, RobLa, Tim, Brad, Chris, Greg, Nik, Antoine, Sam, Aaron, Chad, Dan’s presence in the Google Doc
SUL
- Bug 52206 is fixed! (mystery redirect to wmfwiki)
- Also bug 53498! (mystery redirect to Special:AbuseFilter)
- https://bugzilla.wikimedia.org/show_bug.cgi?id=54513
OAuth
- Please document consequences of changing $wgOAuthSecretKey
- Bug 54716 (UI issue)
- Meeting set up with UTRS developers (1st Oct) and ACC developers (2nd Oct) to help them to test OAuth.
- Dan in contact with Quim about reaching out to other developers too.
OpenID
- Tried to set it up on a labs instance. Filed 7 bugs.
Search
- Approval to go to primary on cawiki
- Haven’t heard on enwikisource
- Oh God, still more bugs
- LVS wrapping Elasticsearch for easier configurability + monitoring
- Done puppet & dns changes, needs ops review
Quarterly review https://www.mediawiki.org/wiki/Site_performance_and_architecture#Deployment_sprint
Other
- https://bugzilla.wikimedia.org/show_bug.cgi?id=53687 - No page history visible of a deleted page on nl.wikipedia
- Tim: unclear whether it is actual data loss or an id glitch.
- https://bugzilla.wikimedia.org/show_bug.cgi?id=54647 - Anything Bryan should do for followup?
- See also: https://www.mediawiki.org/wiki/User:BDavis_(WMF)/Notes/Finding_Files_To_Purge
- Antoine: We had a purge cache issues since August 17th iirc. HTCP purges were not sent to Amsterdam datacenter. I can not remember the details nor when it got fixed though.
- Bryan: Faidon thought it was fixed on 09-26. I think that was based on looking at Ganglia graphs.
- Quoting Faidon right now: “multicast outage that we had from Sep 22nd until the 26th”
- definitely follow up with ops. During european business hours: Mark / Faidon, for SF hours, no clue :( Maybe Leslie (network engineer).
- Bryan need shell access on cluster.
- Antoine to give procedure