14:00:43 <jbernard> #startmeeting cinder 14:00:43 <opendevmeet> Meeting started Wed May 15 14:00:43 2024 UTC and is due to finish in 60 minutes. The chair is jbernard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:43 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:43 <opendevmeet> The meeting name has been set to 'cinder' 14:00:53 <jbernard> #topic roll call 14:01:02 <zaitcev> o/ 14:01:19 <akawai> o/ 14:01:26 <whoami-rajat> hi 14:01:32 <rosmaita> o/ 14:01:47 <eharney> hi 14:02:58 <sp-bmilanov> hello 14:02:59 <crohmann> hi 14:05:07 <jbernard> welcome everyone 14:05:13 <jbernard> #topic annoucements 14:05:39 <jbernard> D-1 milestone is tomorrow 14:06:02 <jbernard> i dont think it means a lot for us, libraries will be released 14:06:14 <jbernard> but otherwise we just keep rolling 14:06:30 <jbernard> please correct me if im missing something 14:06:59 <jbernard> this friday is our festival of reviews 14:07:16 <jbernard> they have been piling up, so there appears to be quite a bit to do 14:07:26 <jbernard> if you can help out, that would great 14:07:33 <jbernard> would *be*, even 14:08:13 <jbernard> that's all ive got 14:08:16 <jbernard> for annoucements 14:08:23 <jbernard> moving onto topics 14:08:35 <jbernard> #topic project-specific milestone dates 14:08:54 <jbernard> I need to define specific dates for us for the following milestones 14:08:59 <jbernard> spec freeze 14:09:04 <jbernard> new driver merged deadline 14:09:06 <jbernard> and 14:09:10 <jbernard> new target driver merge deadline 14:09:41 <jbernard> in looking at last cycle's scheulde, I will likely pick dates consistent with what we've done in the past 14:10:06 <jbernard> does anyone have an specific concerns about deadlines? do we think the default dates will work? 14:11:02 * jungleboyj sneaks in late 14:11:03 <whoami-rajat> it might also be worth looking into the Bobcat cycle deadlines since that aligns better with the release schedule of Dalmatian 14:11:20 <jbernard> whoami-rajat: that's a good idea 14:11:54 <jbernard> ok, i will propose those dates to the list soon and submit them for review if there are no objections 14:12:26 <jbernard> #topic antelope, patch to bump ceph to reef 14:12:39 <jbernard> #link https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/915146 14:13:02 <jbernard> ^ this patch updates ceph from quincy to reef for 2024.1 (antelope) tests 14:13:18 <jbernard> it doesn't have input from us yet 14:13:59 <jbernard> the logic seems reasonable, but i wanted to raise it in case you have an objection 14:14:54 <jbernard> whoami-rajat: has this happened before? i think ideally we'd want to test on both 14:15:39 <rosmaita> well, reef was released in the fall, i think there should be an "s" release happening soon (unless they changed their release schedule) 14:16:33 <whoami-rajat> so this has already merged in Caracal and bobcat and now we are looking into merging it in antelope 14:16:35 <jbernard> i dont think anything has been announced yet 14:16:55 <whoami-rajat> i remember we used to have a compatibility doc between ceph and openstack releases 14:17:20 <rosmaita> ok, i am wrong there ... preliminary release of reef was in the fall, the official release was march 11 14:17:29 <rosmaita> so reef is the latest 14:17:55 <rosmaita> the compatibility doc is the rbd driver docs 14:18:00 <jbernard> i think the logic is that, as time goes on, more folks are likely to be deploying against reef than quincy 14:18:39 <rosmaita> "For a given OpenStack release, Cinder supports the current Ceph active stable releases plus the two prior releases." 14:18:39 <crohmann> yes. We run Reef with Zed 14:19:36 <whoami-rajat> i understand the reasoning of using the latest version in master, just trying to understand if it makes sense for stable branches and till how long we can backport it without causing a breaking change 14:20:00 <jbernard> this is my question as well 14:20:15 <crohmann> I doubt there were client affecting changes for RBD in a while. 14:20:56 <crohmann> The only thing really bugging me are the Ceph caps which are so poorly documented (https://bugs.launchpad.net/nova/+bug/2051244) 14:22:35 <rosmaita> crohmann: have we still not documented that from the cinder side? 14:22:50 <crohmann> nope, no movement to the whole issue. 14:23:27 <rosmaita> what do we still need on that? 14:24:22 <crohmann> We need to agree which caps are ACTUALLY required (which pool by which client with which access level) and also convert this to the profile style of caps that Ceph STRONGLY advises to use 14:27:00 <jbernard> #action movement on ceph caps 14:27:08 <rosmaita> i know you've been working on this for about a year or so 14:27:25 <rosmaita> not sure what we can do to speed things up 14:27:51 <whoami-rajat> can get ask the operators to provide recommended profiles that they are using in their deployment currently? we can make changes to it but at least it will be a good starting point 14:28:01 <whoami-rajat> s/get/we 14:28:03 <rosmaita> other than just post a doc with whatever you think is correct, and let people correct it if they run into a problem 14:28:27 <crohmann> It seems that things have changed at least to profiles in the Documentation at Ceph (https://docs.ceph.com/en/latest/rbd/rbd-openstack/#setup-ceph-client-authentication) 14:28:35 <eharney> i think it'd be useful to add this to devstack-plugin-ceph so we can test it 14:29:03 <crohmann> eharney: Yes, that is one of the raised issues in the referenced Launchpad bug ... that devstack does take shortcuts here 14:29:04 <rosmaita> eharney: ++ 14:29:58 <crohmann> whoami-rajat: My launchpad bug (which I believe you asked me to open) actually DOES contain my own config 14:31:32 <rosmaita> crohmann: are you happy with your config, or do you hope to narrow it some more? 14:32:11 <whoami-rajat> crohmann, okay then as rosmaita said, we can use that as a starting point and review it 14:32:35 <eharney> it could potentially be narrowed more depending on things like whether glance is using ceph as a backend or not 14:32:42 <crohmann> I did not mean to hijack the topics section though. I'd appreciate if someone joins the conversation in the ticket. I also gladly push some changes to docs or otherwise if there is some aggreement on what is "correct" 14:33:30 <rosmaita> i think no one wants to commit to "correct", because it might not work! 14:33:39 <jbernard> crohmann: no worries, this is important 14:33:45 <eharney> this is why i think the most valuable next step is to implement it in the devstack plugin 14:34:39 <jungleboyj> ++ 14:35:15 <crohmann> I'll push a change for devstack as a starting point then. Otherwise ... join the fun train at https://bugs.launchpad.net/nova/+bug/2051244 14:35:56 <rosmaita> that sounds like a good plan 14:36:16 <jungleboyj> Agreed. 14:36:28 <jbernard> excellent 14:36:34 <jbernard> #topic backup patches 14:37:11 <zaitcev> oh no 14:37:12 <jbernard> really quickly, as crohmann notes as well, there are several c-bak patches accumulating, I think we should try to review these as part of our festival efforts on friday 14:37:15 <zaitcev> what about them? 14:37:28 <zaitcev> Totally! 14:37:35 <jbernard> great ;) 14:38:05 <jbernard> that's all for topics I had 14:38:30 <jbernard> anything else from anyone? how are things going for everybody? 14:39:00 <zaitcev> Backup encryption is progressing. Just slowly. 14:40:03 <jbernard> progress is good! 14:40:20 <jbernard> I'll spin through the review requests today 14:40:31 <jbernard> #link https://etherpad.opendev.org/p/cinder-dalmatian-meetings#L104 14:40:46 <jbernard> if anyone would like to raise specific concerns, feel free 14:40:51 <jbernard> #topic open discussion 14:41:14 * sp-bmilanov is trying to figure out why the gate pipeline has failed, twice, on different jobs for 803823, and if he should re-run until all jobs pass 14:41:38 <crohmann> jbernard: I put my change (https://review.opendev.org/c/openstack/cinder/+/810457) in there ... as I have a specific question that I need answer to prior to pushing a hopefully final patchset. See https://review.opendev.org/c/openstack/cinder/+/810457/comment/31a6c2c2_4750ec07/ 14:43:03 <eharney> there was a suggestion in this requirements change that we merge it at the beginning of this cycle: https://review.opendev.org/c/openstack/cinder/+/908193 14:44:59 <jbernard> crohmann: ok 14:45:34 <crohmann> There also is another question about a (potentially broken) test failing in https://review.opendev.org/c/openstack/cinder/+/484729/comments/b6da8f16_5c9481e3?tab=comments, also keeping me from finishing this change for a final review 14:45:42 <rosmaita> sp-bmilanov: i just issued a recheck, looks like a random thing 14:46:42 <whoami-rajat> i see this issue in the ceph tempest job, could it be related to our recent RBD changes 14:46:44 <whoami-rajat> May 15 00:45:03.833658 np0037512504 cinder-volume[113176]: ERROR oslo_messaging.rpc.server rbd.ImageHasSnapshots: [errno 39] RBD image has snapshots (error removing image) 14:46:56 <jbernard> #action movement on crohmann blocking questions 14:47:03 <eharney> whoami-rajat: yeah i was looking at that one earlier. first time i've seen it, not sure why yet -- i think it references a read only image 14:47:48 <whoami-rajat> yeah it could be 14:47:56 <whoami-rajat> May 15 00:45:03.833658 np0037512504 cinder-volume[113176]: ERROR oslo_messaging.rpc.server [req-e5453de1-89c1-41e7-9c97-12b3b82181d3 req-fe8a5e99-6bbe-4cbb-a4ba-ad8f0f86e574 tempest-TestVolumeBootPattern-189137301 None] Exception during message handling: rbd.ReadOnlyImage: [errno 30] RBD read-only image (error flattening b'volume-74ed150e-1809-41ad-a759-d388e5130a3e') 14:48:02 <sp-bmilanov> rosmaita: thanks! 14:49:23 <jbernard> eharney: the moto bump needs to merge nowish? 14:50:08 <eharney> jbernard: i don't know about "needs", but it was bumped out to not be at the end of the last cycle. when is the right time? 14:51:29 <jbernard> eharney: if zuul posts updated results, i dont see why we couldn't get this in now 14:56:01 <jbernard> ok, if there's nothing else, i think we're done 14:57:06 <jbernard> thanks everyone 14:57:09 <jbernard> #endmeeting