[11:14:37] !log admin Merged the osd scheduler change for all osds, applying on all cloudcephosd* (T273791) [11:14:41] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Admin/SAL [11:14:42] T273791: [ceph] Change the io scheduler none/noop - https://phabricator.wikimedia.org/T273791 [12:08:41] !log tools.mabot Update MABot to 9ce43db [12:08:42] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.mabot/SAL [15:23:51] !log admin icinga-downtime for 2h everything *labs *cloud for openstack upgrades [15:23:53] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Admin/SAL [16:58:35] hmm, at least mine has no websvc [16:59:04] 4GiB being the default limit sounds correct [16:59:08] https://wikitech.wikimedia.org/wiki/Help:Toolforge/Grid#Allocating_additional_memory doesn't mention a limit [16:59:28] the exec nodes only have 8GiB each so that's a hard limit [16:59:45] someone try to submit a job for 4.1 GiB and see what happens [17:00:17] the grid engine accounting is not precise at all either, so ... yeah [17:00:36] a limit on a grid job is more of a hope than a guarantee [17:01:14] but I'm pretty sure the scheduler will just stick your job in the wait queue forever if you set a limit higher than the exec nodes have space for [17:01:18] seems like all the legacy :-p [17:05:30] revi: unsatisfying answer is apparently that the limit is the size of an empty exec node's free space as seen by grid engine. The web-memlimit stuff was what I was remembering as "quota". That's really just the ram asked for by `webservice` when it launches a grid job though. [17:05:58] that's satisfying enough, thanks! [19:55:28] !log download rebuilt download server with a cinder-backed volume; now the active host is download-02 [19:55:30] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Download/SAL [19:56:15] !log download deleted the now-unused download-01 [19:56:18] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Download/SAL [20:13:40] hi, i'm trying to copy a big bunch of data between two vps instances. is there a guide / recommended way to do this, or do i need to figure out the commands myself? [20:13:50] (15 GB or so) [20:27:38] MatmaRex: you're in luck! https://techblog.wikimedia.org/2021/02/05/cinder-on-cloud-vps/ [20:27:51] you can move the files onto a cinder volume and then just move that volume between instances [20:28:32] that page says there's a 10 GB quota on that [20:28:40] (i read it already) [20:28:44] true but you can request a quota increase [20:29:13] (it looked scary and i just figured out how to do the LVM thing) [20:30:10] LVM going to get phased out so best to just move to Cinder now — it shouldn't be especially scary and if it is that'll teach me about what features to add [20:32:56] andrewbogott: i want 100 GB of space and i don't want to wait for 7 days :P [20:33:25] (for context, this is for https://patchdemo.wmflabs.org/ , which has apparently become popular enough to start running out of disk space) [20:37:19] andrewbogott: when i click "Create Volume" in horizon, i get a red popup that says "Danger: An error occurred. Please try again later." [20:37:34] that's new! I'll look [20:37:37] i don't really want to experiment with this now, to be honest ;) [20:38:50] and in browser console, i see this: [20:38:51] GET https://horizon.wikimedia.org/project/volumes/create/ 500 [20:43:10] I'm looking at the issue, probably won't be very responsive in the meantime [20:43:21] there was a major version upgrade this morning so it's probably some api versioning thing [20:56:25] MatmaRex: should work now, you'll need to log in again [20:57:33] thanks, it does