[04:28:26] <marostegui>	 I am going to start with the s7 primary master dc switch
[04:50:32] <marostegui>	 primary switched
[05:00:02] <Amir1>	 marostegui: please let me know once you're done with the old s7 master mwhahaha, I call dibs before Arnaud 
[05:00:51] <marostegui>	 He's off till the 10th :p
[05:01:43] <Amir1>	 one time I manage to be earlier
[06:57:29] <marostegui>	 Amir1: db1181 is all yours, please repool it after you are done
[08:37:22] <Amir1>	 Thanks
[13:18:09] <elukey>	 urandom: o/
[13:18:26] <elukey>	 if you have time we can move session store to pki
[13:28:56] <Emperor>	 If one backup store knows about a file and the other doesn't, does that mean it's always only been available in one DC?
[13:30:18] <Emperor>	 Context is T363995
[13:30:19] <stashbot>	 T363995: Commons: File not found - https://phabricator.wikimedia.org/T363995
[13:31:53] <Emperor>	 Neither ms cluster has the object, I have no non-404 accesses to the original in swift logs, not any DELETE (but there is a whole pile of DELETE of the thumbs on May 1 at 23:29:13), backup2001 doesn't know of the object, but backup1001 does
[13:42:27] <urandom>	 elukey: ready if you are
[13:45:20] <Emperor>	 relatedly, should I just put the object back into the correct swift path, or does mw need fettling somehow? I'd ask j.ynus but their OOO. Amir1?
[14:01:44] * Emperor doesn't feel confident enough that just shoving the objects from backups back into swift is the correct approach
[14:21:20] <elukey>	 urandom: sorry I was in a meeting
[14:22:16] <elukey>	 gimme 10 mins and I'll start :)
[14:24:09] <urandom>	 kk