[09:56:41] [02puppet] 07RhinosF1 pushed 031 commit to 03master [+0/-0/±1] 13https://git.io/J8lso [09:56:42] [02puppet] 07RhinosF1 0343d0055 - services: make ssh + ping(4|6) #page [10:01:00] CUSTOM - ping4 #page on bots1 is UNKNOWN: rhinosf1 Test [10:02:00] If #page is in an alert, treat it as important [10:02:46] They'll be directed some place useful soon [10:03:00] Probably by Miraheze bot + hund [10:03:18] Maybe something like slack [16:04:49] [02puppet] 07sario528 pushed 031 commit to 03sario528-patch-1 [+0/-0/±1] 13https://git.io/J8gSk [16:04:51] [02puppet] 07sario528 03f931136 - Add ##Sario [16:04:52] [02puppet] 07sario528 created branch 03sario528-patch-1 - 13https://git.io/JJGvA [16:05:27] [02puppet] 07sario528 opened pull request 03#195: Add ##Sario - 13https://git.io/J8gSi [16:06:08] RhinosF1: got a moment to review? ^ [16:07:00] Sario528: missing a ] [16:07:36] Sario: * [16:08:49] [02puppet] 07sario528 pushed 031 commit to 03sario528-patch-1 [+0/-0/±1] 13https://git.io/J8gHd [16:08:50] [02puppet] 07sario528 03d04e558 - fix [16:08:51] [02puppet] 07sario528 synchronize pull request 03#195: Add ##Sario - 13https://git.io/J8gSi [16:10:33] And Sario528 still fires a notification since it contains the string Sario, it just doesn't get colored [16:13:23] [02puppet] 07RhinosF1 closed pull request 03#195: Add ##Sario - 13https://git.io/J8gSi [16:13:25] [02puppet] 07RhinosF1 pushed 031 commit to 03master [+0/-0/±1] 13https://git.io/J8gdn [16:13:26] [02puppet] 07sario528 03e0dfa04 - Add ##Sario (#195) * Add ##Sario * fix [16:13:28] [02puppet] 07RhinosF1 deleted branch 03sario528-patch-1 - 13https://git.io/JJGvA [16:13:29] [02puppet] 07RhinosF1 deleted branch 03sario528-patch-1 [16:13:40] RhinosF1: Thanks [16:13:40] Sario: ^ [16:13:53] I'll run puppet and test [16:13:57] Don't forget to use the cache reload command [16:14:24] yup [16:21:49] Sario: missing } at end [16:21:54] We should have CI [16:22:02] Ahh, I'll fix [16:24:54] [02puppet] 07sario528 pushed 031 commit to 03sario528-patch-1 [+0/-0/±1] 13https://git.io/J8gju [16:24:55] [02puppet] 07sario528 03eb5ef74 - fix again [16:24:57] [02puppet] 07sario528 created branch 03sario528-patch-1 - 13https://git.io/JJGvA [16:24:58] [02puppet] 07sario528 opened pull request 03#196: fix again - 13https://git.io/J8gj2 [16:27:36] [02puppet] 07RhinosF1 closed pull request 03#196: fix again - 13https://git.io/J8gj2 [16:27:38] [02puppet] 07RhinosF1 pushed 031 commit to 03master [+0/-0/±1] 13https://git.io/J82vI [16:27:39] [02puppet] 07sario528 037b834d6 - fix again (#196) [16:27:41] [02puppet] 07RhinosF1 deleted branch 03sario528-patch-1 - 13https://git.io/JJGvA [16:27:42] [02puppet] 07RhinosF1 deleted branch 03sario528-patch-1 [16:27:47] Sario: merges [16:28:06] puppet running [16:42:40] $log restarting prod [16:42:47] Saved item "restarting prod" [16:44:08] $log restarting beta [16:44:13] PROBLEM - MirahezeBots sopel on bots1 is WARNING: PROCS WARNING: 1 process with command name 'sopel' [16:44:15] Saved item "restarting beta" [16:46:13] RECOVERY - MirahezeBots sopel on bots1 is OK: PROCS OK: 2 processes with command name 'sopel' [16:46:16] $log restart beta again [16:46:22] Saved item "restart beta again" [17:01:52] We need to plan bullseye [17:02:08] It's out in like a week [17:02:19] I'll try and reimage test2001 nearly as soon as possible [17:02:38] Then start provisioning other servers [17:02:44] Mysql failover will be fun [17:02:57] And probably need a fair bit of phab downtime [17:04:05] MacFan4000: I assume mysqldump & then mysql < is best [17:04:36] It might be worth testing latency to the aarch platform [17:04:54] Is it possible to set MariaDB up on test2001 and connect them [17:04:56] yes, though phabricator has multiple databases [17:05:17] MacFan4000: yeah I know it's huge [17:05:40] I just want to see how phab will behave when we have db2002 but are still on tools1 [17:06:05] I have a shell script that can go through and export them all to .sql files - thats how I moved the dbs for testwiki phab [17:06:43] and then I have another script that goes through and imports them all [17:08:15] MacFan4000: okay, can you test what latency is from tools to test [17:08:20] Icinga might know tbh [17:09:20] 130+ [17:09:36] So we're gonna see a pretty big performance regression [17:09:50] At least tools is easy to failover [17:09:55] https://www.irccloud.com/pastebin/UsJfHkyN/ [17:10:02] there's the output of ping [17:10:10] Ye [17:10:31] We can probably fail phab over in one go tbh [17:10:48] You can operate phab from multiple app servers [17:11:16] So it won't be hard to do it in one go [17:12:43] If we setup a read only phab-new that connects to db1 using NAT64 [17:13:14] So image tools2002 first [17:14:11] yeah we can have 2 appservers, so 1st pool tools2002 [17:14:35] Then we can switchover from db1 -> db2002 [17:14:49] Then we depool tools1 [17:14:59] That will work [17:15:34] We should document this plan [17:22:14] Sario: https://phab.mirahezebots.org/T269 [17:22:15] [url] ⚓ T269 Failover phab & db | phab.mirahezebots.org