14:00:26 #startmeeting cinder 14:00:26 Meeting started Wed Dec 8 14:00:26 2021 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:26 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:26 The meeting name has been set to 'cinder' 14:00:28 hi 14:00:35 #topic roll call 14:00:39 o/ 14:00:52 o/ 14:00:55 kind of. 14:01:01 hi 14:01:03 probably a light turnout today 14:01:13 it is a holiday in a surprising number of countries 14:01:22 well, surprising if you are an ugly american like me 14:01:22 slackers... 14:01:51 enriquetaso_: i thought today is a holiday in Argentina? 14:02:05 hi all 14:02:08 hi 14:02:16 it is 14:02:19 :P 14:02:25 enriquetaso_: bonus points for you! 14:02:38 https://www.holidayscalendar.com/ 14:03:03 ok, i think we can get started 14:03:16 argentina it's not in the list simondodsley :P 14:03:22 not sure why 14:03:23 kind of a light agenda, mostly announcements, but that will leave plenty of time for open discussion 14:03:32 #topic announcements 14:03:50 first, you probably noticed that we are not holding this meeting in video right now 14:04:04 we'll have the followup session for the midcycle tomorrow 14:04:14 details here: 14:04:17 #link http://lists.openstack.org/pipermail/openstack-discuss/2021-December/026196.html 14:04:29 and, there is the etherpad also 14:04:39 #link https://etherpad.opendev.org/p/cinder-yoga-midcycles 14:04:51 time will be 1400-1430 UTC 14:05:21 topic is the new quotas work, and if there's time, also the community goal ("secure and consistent RBAC") 14:05:29 it will be recorded if you can't make it 14:05:53 next, early Yoga releases have happened 14:06:02 os-brick: 5.1.0 14:06:08 python-cinderclient: 8.2.0 14:06:30 i think neither really needed a release, but the release team likes early releases in the cycle to shake out problems 14:06:52 next, our cycle-trailing deliverable must be released next week 14:06:55 that's cinderlib 14:07:13 there is one patch that has to merge immediately, so cores, please look it over: 14:07:22 #link https://review.opendev.org/c/openstack/cinderlib/+/820391 14:07:28 it's a simple patch 14:08:00 just noticed that i didn't put a link in the commit message to the docs explaining why we have to do this 14:08:10 so i will put it here (soon as i find it) 14:08:35 #link https://docs.openstack.org/cinderlib/latest/contributor/contributing.html#cinderlib-releases 14:09:03 looking at the open patches, i don't see anything else that needs to get in to the xena release 14:09:16 #link https://review.opendev.org/q/project:openstack/cinderlib+branch:master+status:open 14:09:59 so https://review.opendev.org/c/openstack/cinderlib/+/820391 doesn't get forgotten, can I have 2 cores commit to looking at it today? 14:10:49 820391 has +2 already? 14:11:13 ok, that was fast 14:11:18 thanks e0ne and eharney 14:11:29 rosmaita: np 14:11:38 just needs the workflow vote 14:11:42 i think there was a race condition there 14:12:01 thanks! 14:12:20 the release team already has a patch up for cinderlib release, so we're all set there 14:12:42 next reminder: 14:12:48 spec freeze is next week 14:12:57 #link https://releases.openstack.org/yoga/schedule.html#y-cinder-spec-freeze 14:13:29 i tried to get comments on all the proposed specs, but there are a few i still have drafts on 14:13:32 #link https://review.opendev.org/q/project:openstack/cinderlib+branch:master+status:open 14:13:46 i'll get those done today, and then start re-reviewing 14:14:19 also next week is the monthly Festival of XS Reviews 14:14:29 1400-1600 UTC on Friday 14:14:47 next week is a pretty busy week! 14:14:57 something a bit further out is the new driver merge deadline 14:15:02 which is 21 January 14:15:09 #link https://releases.openstack.org/yoga/schedule.html#cinder-new-driver-merge-deadline 14:15:28 i know people are itching for reviews on their new driver patches 14:15:45 but specs reviews are the priority over the next week 14:15:48 in the meantime 14:16:01 you can look through our handy new driver reviewer checklist 14:16:11 and make sure you have all those things addressed 14:16:21 #link https://docs.openstack.org/cinder/latest/contributor/new_driver_checklist.html 14:16:45 and, of course, make sure your third party CI is working 14:16:56 and displaying results in an easily consumable format 14:17:23 that is, do not require people to download and decompress an archive file in order to see the logs etc 14:17:39 ok, that's all for announcements 14:17:55 #topic TC collection of service pain points 14:18:33 the TC has been collecting operator feedback with the idea that there could be a service-specific community goal to fix a pain point 14:18:48 the info has been collected here: 14:18:57 #link https://etherpad.opendev.org/p/pain-point-elimination 14:19:14 there is a video meeting to discuss this today, right after our meeting 14:19:23 #link http://lists.openstack.org/pipermail/openstack-discuss/2021-December/026207.html 14:19:40 i will attend, though i'm not exactly sure what it's about 14:19:50 anyway, there are 2 items currently listed for cinder 14:20:01 1. volume backup rotation 14:20:12 2. cinder backup fails due to user token expiration with swift backend 14:20:39 rosmaita: What it is about will hopefully be clearer after the discussion. 14:20:51 jungleboyj: hope so! 14:21:20 Goal was to look at some of the higher level problems that our users have and if something can be done community wide to help address those items. 14:21:44 ok, that's good from a cinder point of view then, because these are both minor issues 14:21:56 ++ 14:23:48 the first one looks like a minor enhancement, unless the idea is that cinder should store the number of backups, then it would require some thought about whether it's per-volume or per-project or what 14:24:28 the second one has an alternative fix (if it is in fact a problem) 14:25:14 short story is, the status of keystone "trusts" is kind of suspect, and the advice i got from the keystone team is don't use them in new code 14:26:09 #topic openstackclient feature parity 14:26:17 this one is a surprise 14:26:29 #link http://lists.openstack.org/pipermail/openstack-discuss/2021-December/026192.html 14:26:49 the claim is made there that openstack SDK fully supports cinder 14:27:00 Whaaaaat!?! 14:27:03 yeah 14:27:14 i haven't had time to look closely at that 14:27:22 Well, ship it. 14:27:26 ;-) 14:27:57 hmm 14:28:47 the idea is that if SDK fully supports Block Storage API v3, then since the osc uses SDK, it's just a small matter of time before the osc will fully support it too 14:29:11 Well, that would be awesome if true. 14:30:15 #link https://opendev.org/openstack/openstacksdk/src/branch/master/openstack/block_storage 14:30:39 jungleboyj: +1 14:31:22 looking at the history makes me kind of skeptical 14:31:44 #link https://opendev.org/openstack/openstacksdk/commits/branch/master/openstack/block_storage 14:33:00 if you look at the email i linked above, this came up because a pain point is "why can't i use the openstackclient for everything?" 14:34:40 ok, that's it from me ... let's have open discussion 14:34:46 #topic open discussion 14:36:20 ... or not 14:38:10 enriquetaso_: thank you for your weekly bug report: 14:38:13 #link http://lists.openstack.org/pipermail/openstack-discuss/2021-December/026225.html 14:38:14 i put up a couple of patches a bit ago that just remove dead code, maybe not a tough review if anyone wants to take a look: https://review.opendev.org/c/openstack/cinder/+/800703 14:38:47 I have only one question regarding the bug report 14:38:59 sure 14:39:03 maybe i can do it in the open discussion and skip the bug meeting 14:39:10 sounds good to me! 14:39:24 bug_1 cinder schedules backups on disabled backup services 14:39:24 #link https://bugs.launchpad.net/cinder/+bug/1952804 14:39:35 Summary: Set up multiple cinder-backup nodes then mark one of them as disabled via 'openstack volume service set --disable'. Create some backups and those backups get sent to the disabled node. The reporter points out that the real issue here is the lack of any kind of feedback during the backup; ideally there would be some way to detect stuck jobs and retry them. 14:39:44 Maybe this is too obvious but I'm not sure what to reply to the reporter here. 14:39:45 The creation of the backup depends on the backend you're using, if it takes too long to create the backup there's not much we can do excerpts for a retry. However, cinder shouldn't use the disable nodes for backups. 14:45:38 ..guess I'll just mark the bug as medium importance and we can fix it :P 14:46:38 it will take some investigating to figure out what to do with this, there may be some room to improve usability w/ async error messages 14:47:47 yeah, i think a related issue is 'disabled' temporarily (fine for jobs to get picked up on restart) vs 'disabled permanently', i.e., the node is never coming back 14:50:20 sure 14:50:35 OK, that's all I have 14:50:45 14 bugs reported in two weeks!! 14:50:45 The good side it's that most of them have fixes proposed on master but please take a look at the full list. 14:51:07 thanks enriquetaso_ ! 14:51:38 anyone else have anything to discuss? 14:52:15 ok, as a reminder, all cinder-cores need to be looking at specs over the next week 14:52:46 i guess that's all for today 14:52:53 thanks everyone, and have a good rest of the week 14:53:04 don't forget the midcycle followup session tomorrow at 1400 UTC 14:54:52 all right, we can close the meeting 14:54:54 #endmeeting