[06:21:08] for s7 the extlinks clean up drops 70GB per host https://grafana.wikimedia.org/d/000000377/host-overview?orgId=1&var-server=db2120&var-datasource=thanos&var-cluster=wmcs&from=1692854589441&to=1692858022299&viewPanel=28 [06:34:20] marostegui: db2179 candidate master of s4 in codfw is in api group, should I remove it with dbctl edit? [06:44:46] no need [06:45:02] candidate masters do have other roles too [06:45:05] it is fine [06:45:32] in the switchover checklist there's an item to review those things when a master swap is done [07:09:04] okay makes sense [07:34:17] FYI https://phabricator.wikimedia.org/T335080#9115867 [07:52:11] thanks for taking care of it [08:06:15] Also I am going to start failing over all the misc proxies too [08:06:28] m5 can be fun [09:10:08] I did s4 in eqiad and codfw this morning [12:48:32] moritzm: dns1005 fingerprint isn't here? https://wikitech.wikimedia.org/wiki/Help:SSH_Fingerprints is it somewhere else? [12:59:43] see https://config-master.wikimedia.org/known_hosts [13:00:10] the wikitech is a legacy mechanism which is only being done for bastions since some people were asking for it to continue [13:00:19] ah cool thanks [19:41:24] anybody know if user 'querysampler' still exists in the wiki replicas? i'm able to login with my toolforge db user (-u u2962) on clouddb1017 via /var/run/mysqld/mysqld.s1.sock as well as TCP localhost:3311, but no luck with -u querysampler using the creds in clouddb-wikireplicas-query-1.clouddb-services.eqiad1.wikimedia.cloud:/etc/querysampler-config.yaml . the wiki-replicas.sql suggests login from %. [19:43:13] but last data collection seems back from 2021, too, so i figured it could have been clobbered since then. i was just going to run the script with a manual override to store in a different .db file than the one the service was using back then. [19:48:50] dr0ptp4kt: i don't think we have that user anymore but I can check tomorrow [20:09:11] thanks! i'm sorry if this ended up pinging you so late in the day.