[10:57:35] So, here it is a tip for database requests: If all of them are priority:high, "high" just becomes "normal" [10:57:39] :-) [11:04:49] jynus: where to triage tickets fall in on that scale? :) [11:05:17] ^sorry, didn't get that [11:05:42] Never mind :) [11:06:06] gotcha, well, technically, being this week on clinic duty, it should be "high" this week [11:06:12] :-) [11:07:24] there are some workflow issues, globally, not only only in Databases or ops, will try to improve that [19:54:25] jynus, hey, https://wikitech.wikimedia.org/wiki/Schema_changes/Coordination_proposal seems fine but I wonder what you mean by "created by developers, deployed by the deployment team, and over-sighted by operations and performance teams, without a hard dependency on DBAs, like any other piece of code, but that is not technically possible now." [19:55:12] not technically possible because developers typically aren't trained on how to do these properly or don't have the relevant mysql permissions? [19:55:18] that is not part of the short term proposla [19:55:25] that is a wich for the future [19:55:31] yes [19:55:34] *wish [19:55:37] I understand that [19:56:02] it it technically possible to deploy schema changes as we do deploy code [19:56:05] with a tool [19:56:13] we just do not have that tool [19:56:17] ah. [19:56:20] okay [19:56:27] we hace scripts that I use [19:56:39] checking every case separatelly [19:57:03] chosing if using one or another script and praying that things do not break [19:57:24] it should be possible, with some engineering, make a "schema scap" [19:57:52] it is just my excuse for, "sorry, this is what we have now" [19:58:38] LGTM, which is better than the false documentation than we have now [19:59:42] maybe not evend doing any online change, just a switchover- but thare are so many blockers right now! [20:00:16] row based replication, GTIDs, proxying