[07:01:53] Hi, I have a project that is impacted by the Debian Bullseye migration. I am working on hopefully moving the project to Toolforge, but I will need some time (probably up to two to three weeks) as I just got my quota increase approved. I am wondering on how to request an extension or exemption? If it turns out that it isn't feasible to move to [07:01:54] Toolforge, I intend to follow the OS migration. The "[REMINDER]: Migrate Cloud VPS Instances from Debian Bullseye" email I got points me here. [07:19:44] !help (apologies if I shouldn't have done that) [07:19:44] If you don't get a response in 15-30 minutes, please email the cloud@ mailing list -- https://wikitech.wikimedia.org/wiki/Help:Cloud_Services_communication [07:47:19] Leaderboard: can you open a phabricator task requesting the extension? We can then follow up there. [07:50:40] OK thanks [12:20:57] working for me now -- thanks dcaro ! [14:58:20] cannot shh to bastions... [14:58:54] godog: ^ you doing reboots? [14:59:50] dcaro: no [14:59:53] ack, looking [15:01:33] but yeah looks like bastion-15 is busted [15:01:36] 14:54 -wmcs-alerts:#wikimedia-cloud-feed- FIRING: InstanceDown: Project tools instance tools-bastion-15 is down - https://prometheus-alerts.wmcloud.org/?q=alertname%3DInstanceDown [15:02:19] https://www.irccloud.com/pastebin/jyTSUMPQ/ [15:02:34] that's the last logs in the console, looking into dmesg (not sure what timestamp is that though xd) [15:02:53] I think it's just out of memory [15:03:00] https://www.irccloud.com/pastebin/UJGGCP12/ [15:03:21] let me check what's using it and I'll kill or reboot the VM [15:03:47] hmm... OOMKiller should be doing it's thing though [15:03:49] ack [15:04:11] too unresponsive, let me reboot [15:04:29] !log tools reboot tools-bastion-15, getting out of memory [15:04:31] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [15:10:56] !log tools hard reboot tools-bastion-15 [15:10:58] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [15:11:41] host is back [15:14:48] Thanks man [15:15:10] sure np! [15:33:18] Question, is there an easier way to reimage an VPS instance besides deleting and recreating? [15:37:47] I think delete-and-recreate is the standard (VMs are cattle not pets) [15:42:07] I figured, thanks [15:43:10] We have to do the same thing for InternetArchiveBot and we're at quota 😬 [15:43:21] (You can request temporary quota bursts) [15:43:49] that sounds fun... lol [15:43:54] luckily I'm not even close to quota [15:47:58] create the new machine before deleting. rsync all the stuff between them. only then delete old instance? [15:50:43] thanks [15:52:37] if you don't use puppet and want to transfer "which packages are installed" you can also do this: dpkg --get-selections > somefile.txt | copy the file to new machine | dpkg --set-selections < somefile.txt and BOOM all the same packages are installed [15:52:51] of course the best fix is to use configuration management like puppet [15:58:08] yeah this time around I think I'll definitely setup a project-level puppet, thanks for the help mutante [16:23:26] yw, happy to answer questions about that if needed [17:45:20] Zppix: using volumes for storing data is really helpful too. a volume can be detached from your N-1 instance and reattached the the new N instance with a couple of clicks in Horizon. Much easier IMO than rsync or other copy mechanisms. [18:00:54] bd808: that very pattern has saved me so much time on my own cloud servers [20:47:59] !log tools.wikibugs Updated channels.yaml to: cab24068909455ccc3a11fb58066a1835d53c4024e9a32574b3894a2bcc3f767 [20:48:02] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.wikibugs/SAL