[00:42:37] (SessionStoreOnNonDedicatedHost) firing: (2) Sessionstore k8s pods are running on non-dedicated hosts - TODO - TODO - https://alerts.wikimedia.org/?q=alertname%3DSessionStoreOnNonDedicatedHost [04:42:37] (SessionStoreOnNonDedicatedHost) firing: (2) Sessionstore k8s pods are running on non-dedicated hosts - TODO - TODO - https://alerts.wikimedia.org/?q=alertname%3DSessionStoreOnNonDedicatedHost [06:22:34] I have switched es5, now going for pc1 [06:59:22] pc1 done [06:59:24] now going for s2 [08:42:38] (SessionStoreOnNonDedicatedHost) firing: (2) Sessionstore k8s pods are running on non-dedicated hosts - TODO - TODO - https://alerts.wikimedia.org/?q=alertname%3DSessionStoreOnNonDedicatedHost [08:45:11] I have ack'ed that alert for now [08:46:01] I gave it 24h [08:59:06] I've asked about it in #wikimedia-serviceops [09:11:53] thank you Emperor [09:48:40] I've documented the "brief interruption ok" and "pause mediabackups" on the spreadsheet (I didn't know there was one) [09:54:19] thanks jynus [09:54:26] that's why I pinged you there :) [09:54:50] I also posted it on the weekly meeting agenda for next monday [10:00:01] volans: https://phabricator.wikimedia.org/P55975#226798 [10:06:29] nice! [M ZK[M#ZK [10:06:35] sorry [10:06:51] the old ones will be garbage collected in time automatically? [10:09:16] they keep the same system as all backups (retention of 3 months) [10:10:09] it is just they have a different name (because the name depends on the policy- something I would like to change at some point) [10:13:46] ok great, thanks [10:14:18] See: https://phabricator.wikimedia.org/P55975#226799 [10:14:59] the problem being that it was an incremental, and why I prefered to do fulls for packages [10:15:23] and it is already saving space [10:19:13] nice! [11:59:18] marostegui: can I play with the old s2 master? [12:03:59] it is pooled [12:04:03] so it is all yours if you depool it [12:27:30] sure [14:56:32] If someone's a moment to look at https://gerrit.wikimedia.org/r/c/operations/puppet/+/995046 I'd be grateful, please :) [14:59:46] done [15:07:51] Thanks :) [16:25:24] Hello! I'd like to run a `CREATE INDEX` on db1195/m2-master.eqiad.wmnet in the otrs database. The table is 28m rows, and running on a test instance takes approximately 50 seconds (slightly smaller table there, so I'd expect 60-70 seconds here). The create statement is `CREATE INDEX ticket_history_change_by ON ticket_history (change_by);`, and the context can be found here: https://phabricator.wikimedia.org/T356404 [16:25:43] I spoke about this with marostegui a few minutes ago, so if there's no objections in ~5 minutes I'll go ahead with this [16:25:48] Go for it