14:00:31 <rosmaita> #startmeeting cinder 14:00:32 <openstack> Meeting started Wed Sep 9 14:00:31 2020 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:35 <openstack> The meeting name has been set to 'cinder' 14:00:42 <rosmaita> #topic roll call 14:00:43 <jungleboyj> o/ Kind of ... Double booked. 14:00:45 <tosky> o/ 14:00:48 <walshh_> hi 14:00:48 <whoami-rajat__> Hi 14:00:49 <kaisers_> hi 14:00:51 <LiangFang> hi 14:00:59 <rosmaita> jungleboyj: that's fine ... you are back, which is the key thing 14:01:05 <e0ne> hi 14:01:12 <sfernand> hi 14:01:17 <jungleboyj> :-) Yeah, still trying to catch up on e-mail. :-) 14:01:38 <rosmaita> good turnout 14:01:46 <enriquetaso> hi 14:01:51 <rosmaita> #link https://etherpad.opendev.org/p/cinder-victoria-meetings 14:02:08 <lseki> o/ 14:02:51 <dviroel> hi 14:02:54 <rosmaita> ok, let's get started 14:03:11 <rosmaita> #topic updates - check job failures update 14:03:41 <rosmaita> there was a code change to tempest on 4 september that broke us 14:03:53 <rosmaita> particularly cinder-tempest-plugin-lvm-lio-barbican and os-brick-src-tempest-lvm-lio-barbican 14:04:06 <rosmaita> you may have seen 3 VolumesBackupsTest failures with "Failed validating 'type' in schema['properties']['backup']['properties']['name']" 14:04:20 <rosmaita> the fix was merged late yesterday 14:04:27 <rosmaita> https://review.opendev.org/#/c/750360/ 14:04:27 <rajinir> o/ 14:04:28 <LiangFang> this blocked us a lot, thanks rosmaita 14:04:43 <rosmaita> took it 11 hours to get checked, gated, and merged 14:05:13 <rosmaita> so go ahead and recheck if you're seeing that problem, should be ok now 14:05:24 <rosmaita> do it fast before something else breaks 14:05:38 <rosmaita> #topic updates - python-cinderclient 14:05:48 <rosmaita> victoria release of cinderclient must happen tomorrow 14:05:57 <rosmaita> one un-merged approved patch: https://review.opendev.org/#/c/739223/ 14:06:14 <rosmaita> depends on this un-merged cinder change: 14:06:14 <rosmaita> #link https://review.opendev.org/#/c/739223/ 14:06:33 <rosmaita> whoami-rajat__: have you had a chance to look at the unit test failures? 14:06:49 <whoami-rajat__> rosmaita: yeah, still trying to figure out if it's a legit error or not 14:06:59 <rosmaita> ok 14:07:02 <whoami-rajat__> nothing is changed along the lines of failure 14:07:08 <smcginnis> #link https://review.opendev.org/#/c/737707 14:07:15 <smcginnis> I think that's the one you meant for the last one. 14:07:25 <whoami-rajat__> yeah ^^ 14:07:28 <rosmaita> smcginnis: ty 14:07:40 <rosmaita> sean is correct, the cinder patch is ^^ 14:08:33 <rosmaita> my only concern is around the policy stuff, but i think that could be addressed in a follow-up 14:08:50 <rosmaita> as far as the stuff that would impact the cinderclient, i think we are ok 14:08:51 <whoami-rajat__> yep working on that as well 14:09:01 <rosmaita> so as long as the check agrees, we should be good 14:09:28 <rosmaita> #topic updates - python-brick-cinderclient-ext 14:09:41 <rosmaita> this also gets its victoria release tomorrow 14:09:48 <rosmaita> not much activity on it recently 14:09:55 <rosmaita> and no open reviews 14:10:14 <rosmaita> so i just wanted to flag that in case anyone knows of some important pending issue? 14:11:05 <rosmaita> let me know, otherwise we will cut the stable/victoria branch at the current HEAD of master and release 14:11:40 <rosmaita> #updates - bugfix protocol reminder 14:11:58 <rosmaita> #topic updates - bugfix protocol reminder 14:12:19 <rosmaita> reminder that any bugfixes for brick or clients go to master first, then must be backported to stable/victoria 14:12:32 <rosmaita> we already have some for brick, will discuss in a minute 14:12:50 <rosmaita> #topic updates - feature freeze 14:12:58 <rosmaita> M-3 release day is FEATURE FREEZE 14:13:05 <rosmaita> this is the current features list: 14:13:14 <rosmaita> https://blueprints.launchpad.net/cinder/victoria 14:13:36 <rosmaita> some are sitting with one +2, so cores, please take a look 14:13:55 <rosmaita> #topic updates - soft string freeze 14:14:06 <rosmaita> M-3 release day is Soft String Freeze 14:14:12 <rosmaita> what this is: 14:14:22 <rosmaita> "You are no longer allowed to accept proposed changes containing modifications in user-facing strings. Such changes should be rejected by the review team and postponed until the next series development opens (which should happen when RC1 is published)." 14:14:40 <rosmaita> so reviewers, please keep that in mind 14:14:54 <rosmaita> the idea is to give the translations team time to get stuff translated 14:15:20 <rosmaita> ok, so that's all the updates 14:15:30 <rosmaita> #topic M-3 release? 14:15:53 <rosmaita> we are not required to do an M-3 release 14:16:42 <rosmaita> so i am not sure whether there's any advantage in doing it 14:16:57 <rosmaita> other than it gives us a hard timestamp for M-3 14:17:44 <rosmaita> i think that all the features with blueprints are underway to the point were everything could merge by Friday 14:17:56 <rosmaita> (given active reviewing and revising activity) 14:18:24 <rosmaita> smcginnis: jungleboyj: what do you think? 14:18:55 <smcginnis> If we don't have anyone downstream asking for a beta release, I think we should skip it. 14:19:00 <rosmaita> i think if we release M-3 it has to be on thursday, wereas if we don't we can declare M-3 to be EOD eastern time on Friday 14:19:02 <jungleboyj> If we aren't required I don't think we have to. I defer to smcginnis . 14:19:21 <rosmaita> ok, that is my feeling ... we are getting packaging feedback from the RDO team 14:19:34 <rosmaita> that's really the only reason i can think of for having a beta 14:19:41 <rosmaita> give packagers a head start 14:19:55 <rosmaita> but i think most wait for RC-1 to see that we are serious about what will be in the release 14:20:11 <smcginnis> Did the RDO team ask for an actual release to help with that? 14:20:11 <e0ne> rosmaita: +1 14:20:30 <rosmaita> smcginnis: no, they are continually doing it themselves 14:20:52 <smcginnis> OK, good. Then I think we can just skip it and get ready for the RC1 release coming up. 14:20:55 <rosmaita> what i meant was, they don't need a beta, and they give us the feedback i think a beta would get us 14:21:00 <jungleboyj> smcginnis: ++ 14:21:04 <rosmaita> ok, cool 14:21:45 <jungleboyj> Sounds good. 14:21:47 <rosmaita> so to be clear: we will declare M-3 to be 5 pm New York time on Friday 14:21:58 <jungleboyj> ++ 14:22:05 <rosmaita> any features not merged by then will require a Feature Freeze Exception 14:22:13 <smcginnis> ++ 14:22:37 <rosmaita> but let's push and get this done, and no FFEs will be necessary and we can concentrate on bug fixes for RC-time 14:22:47 <rosmaita> #topic os-brick situation 14:23:10 <rosmaita> os-brick victoria release, 4.0.0 was released last week 14:23:38 <rosmaita> RDO project was looking at packaging it and noticed that we added a big honking bindep 14:24:14 <smcginnis> Glad they caught that. We didn't in our code reviews. 14:24:16 <rosmaita> which was not good, so with some discussion in openstack-cinder channel yesterday (i think it was), we reverted it 14:24:30 <rosmaita> #link https://review.opendev.org/#/c/750444/1 14:24:43 <rosmaita> and we replaced it with a lightweight fix that is in the gate now: 14:24:52 <rosmaita> #link https://review.opendev.org/#/c/750452/ 14:25:21 <rosmaita> so apologies to smcginnis and hemna, who has actually suggested the lightweight fix 14:25:23 <e0ne> rosmaita: can we release 4.0.1 for victoria with the fix? 14:25:28 <whoami-rajat__> I'm still thinking if the new patch is tested properly and we might end up facing another issue with it 14:25:42 <rosmaita> e0ne: that tis the plan 14:26:03 <rosmaita> whoami-rajat__: can you articulate your concerns a bit? 14:26:26 <rosmaita> so before we can release 4.0.1, we need to backport the changes to stable/victoria 14:26:37 <rosmaita> the changes are: 14:26:38 <whoami-rajat__> i need to collect some links, can i share after the meeting? 14:26:49 <rosmaita> #link https://review.opendev.org/#/c/750649/ 14:26:58 <rosmaita> #link https://review.opendev.org/750655 14:27:12 <rosmaita> whoami-rajat__: that is fine, we can talk in openstack cinder 14:27:26 <rosmaita> ok, so my opinion is: 14:27:30 <whoami-rajat__> ok 14:27:39 <rosmaita> we should definitely revert the change and release 4.0.1 14:27:55 <e0ne> rosmaita: +1 14:28:04 <rosmaita> whether 4.0.1 includes the new fix will depend on whoami-rajat__ and others' discussion in the cinder channel today 14:28:09 <rosmaita> sound good? 14:28:22 <rosmaita> that will at least remove the bad bindep 14:28:49 <rosmaita> but i agree with whoami-rajat__ that we don't want to do another release with more problems 14:29:24 <rosmaita> ok, so once we have 4.0.1 i will file an FFE with the requirements team 14:29:53 <rosmaita> so that we can change the victoria os-brick upper-constraint to 4.0.1 before release 14:30:25 <smcginnis> At least we have a very good reason for seeking an exception. 14:30:55 <jungleboyj> :-) 14:30:59 <rosmaita> ok, so let's have a disucssion in #openstack-cinder after this meeting and sort out what will be in 4.0.1 14:31:03 <rosmaita> thanks! 14:31:11 <rosmaita> #topic volume-local-cache 14:31:27 <rosmaita> LiangFang: you have been dropping in and out, so i guess a bad connection 14:31:37 <LiangFang> yes, now in 14:31:40 <rosmaita> so the current status is 14:31:43 <LiangFang> :) 14:31:52 <rosmaita> os-brick support has been merged into victoria 14:32:03 <rosmaita> the cinder extra-specs patch looks good 14:32:20 <rosmaita> could you give us a quick idea of where the nova changes and CI stuff stand? 14:33:18 <LiangFang> the UT of nova patch previously blocked by os-brick, but now it should work 14:33:50 <LiangFang> but i need to change os-brick from 3.1.0 to 4.0.0 14:33:56 <LiangFang> in nova 14:34:06 <rosmaita> what are the odds of the nova side stuff merging in victoria? 14:35:38 <LiangFang> I have not talked too much with them recently, long time ago, they said want to see storage patch go first. 14:35:52 <rosmaita> right, that makes sense 14:36:26 <rosmaita> smcginnis: jungleboyj: i will need your feedback on what i am about to say 14:36:38 <rosmaita> i think it makes sense not to hold up the extra-specs support 14:36:52 <rosmaita> but also we will not make a big deal about it 14:37:01 <rosmaita> because it doesn't actually do anything useful yet 14:38:29 <jungleboyj> I think I agree with what you have said. :-) 14:38:56 <rosmaita> ok, and i think hemna agrees too (he just +2d the patch) 14:39:05 <jungleboyj> :-) 14:39:17 <rosmaita> ok, and LiangFang we will expect documentation for this in wallaby 14:39:35 <rosmaita> and then we can announce it as a real feature once the nova side is complete in wallaby 14:39:36 <jungleboyj> ++ 14:40:02 <smcginnis> 👍 14:40:29 <LiangFang> I have finished doc and push the patch this week 14:40:40 <jungleboyj> Cool. 14:40:42 <LiangFang> because customer also need the doc 14:40:42 <rosmaita> LiangFang: excellent 14:41:03 <rosmaita> ok, so that is also useful info, that you actually have all this working somewhere 14:41:10 <rosmaita> makes me feel better :) 14:41:23 <rosmaita> #topic selection of forum submissions 14:41:24 <LiangFang> customer want write-back mode 14:41:40 <rosmaita> so we have gathered our brainstorming on this etherpad: 14:41:52 <rosmaita> #link https://etherpad.opendev.org/p/2020-Wallaby-cinder-brainstorming 14:42:12 <rosmaita> so it is now our difficult task to narrow that stuff down to 2 topics 14:42:24 <rosmaita> mission accomplished! 14:42:40 <lseki> lol 14:42:43 <jungleboyj> :-) Done! 14:42:44 <rosmaita> unless anyone absolutely hates one of those, i will submit both of those before the deatline 14:42:49 <rosmaita> *deadline 14:43:09 <jungleboyj> I think those are good topics. 14:43:16 <rosmaita> works for me! 14:43:24 <rosmaita> that was easy 14:43:44 <rosmaita> #topic Should we be able to create a volume from a source with a different volume type? 14:43:54 <rosmaita> #link https://bugs.launchpad.net/cinder/+bug/1874076 14:43:55 <openstack> Launchpad bug 1874076 in Cinder "create volume from source volume fails if different volume type" [Undecided,New] 14:43:58 <rosmaita> walshh_: that's you 14:44:05 <walshh_> thanks 14:44:19 <walshh_> Just wondering if this functionality should be working 14:44:28 <walshh_> it is blocked in the scheduler for us 14:44:46 <walshh_> has anyone here information on this? 14:44:50 <smcginnis> Does seem like it should be an implicit retype. Though that introduces a lot of tricky corner cases. 14:44:52 <jungleboyj> I think that is working as designed. 14:45:13 <smcginnis> Yeah, I think that is intentional. 14:45:17 <walshh_> so you should not be able to create a volume from a source using a different volume type? 14:45:33 <smcginnis> Correct. 14:45:37 <walshh_> ok. that was easy :-) 14:45:37 <jungleboyj> Right, in case the volume is of that type for a reason. 14:45:53 <smcginnis> Workaround would be to create it as the same type, then retype. Or at least attempt to retype. 14:46:13 <jungleboyj> ++ 14:46:24 <walshh_> thanks, i just wanted to check it wasn't just us 14:46:41 <jungleboyj> Not just you. :-) 14:46:52 <walshh_> it is possible to choose a different volume type in the UI 14:46:57 <walshh_> should this be blocked? 14:47:39 <enriquetaso> should we close bug 1874076 then? 14:47:40 <openstack> bug 1874076 in Cinder "create volume from source volume fails if different volume type" [Undecided,New] https://launchpad.net/bugs/1874076 14:48:16 <jungleboyj> It seems that the UI should not allow selecting the destination type. 14:49:19 <smcginnis> Yeah, that might be a horizon bug. 14:49:51 <walshh_> thanks. I will open a bug against horizon. 14:50:12 <walshh_> thanks for the clarification 14:50:37 <jungleboyj> ++ 14:50:40 <jungleboyj> Thanks walshh_ 14:50:51 <rosmaita> i think the problem is that they are just making a general volume-create call, and don't have built-in logic to sort out what parameters can be successfully combined in the call 14:51:14 <rosmaita> our api doesn't help much beyond rejecting your request 14:51:49 <walshh_> it doesn't get beyond the scheduler 14:52:59 <rosmaita> anything else about this topic? 14:53:21 <jungleboyj> So, I think the original bug can be closed as Working as Designed. 14:54:26 <rosmaita> #topic open discussion 14:55:11 <rosmaita> ok, don't forget, after this meeting, discussion in #openstack-cinder about whether the "new" fix for ceph octopus should go into 4.0.1 14:55:16 <rosmaita> (for os-brick) 14:55:29 <tosky> as part of the "move to focal" goal, a few base nodesets are going to switch soon; testing has been done, but still, keep an eye on the jobs in the next soon 14:56:16 <tosky> as part of the "native zuulv3" goal, there is just one job pending, which is the trickiest to adapt, but I will come out with a solution before rc 14:56:19 <tosky> EOL 14:56:29 <rosmaita> tosky: ++ 14:56:33 <tosky> uhm, rather EOF 14:56:44 <rosmaita> thanks for all your work on the zuulv3 nativity 14:58:47 <rosmaita> 1.5 minutes 14:59:28 <rosmaita> ok, thanks everyone ... please review anything that needs it in https://blueprints.launchpad.net/cinder/victoria 14:59:36 <rosmaita> that is the highest priority for reviewing 14:59:50 <rosmaita> #endmeeting