14:00:04 #startmeeting cinder 14:00:04 Meeting started Wed Nov 16 14:00:04 2022 UTC and is due to finish in 60 minutes. The chair is whoami-rajat. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:04 The meeting name has been set to 'cinder' 14:00:06 #topic roll call 14:00:33 o/ 14:00:38 hi 14:01:03 hi 14:01:52 o/ 14:02:39 o/ 14:02:58 #link https://etherpad.opendev.org/p/cinder-antelope-meetings 14:03:00 hi 14:03:42 yough 14:04:57 reasonable attendance, let's get started 14:05:03 #topic announcements 14:05:12 first, Cinder festival of XS reviews this friday 14:05:19 #link https://etherpad.opendev.org/p/cinder-festival-of-reviews 14:05:34 last time it was in PTG week and we didn't get enough time to do it 14:05:53 but since this week is the third friday of the month, we will have festival of XS reviews 14:06:24 I will send a mail later today or tomorrow for a reminder 14:06:30 next, Add topics for midcycle 14:06:39 #link https://etherpad.opendev.org/p/cinder-antelope-midcycles 14:06:56 we are going to have midcycle 1 on 30th Nov which is next to next week 14:07:07 so it's close, so please add your topics to the etherpad ^ 14:07:34 next, Migrating devstack jobs to Jammy (Ubuntu LTS 22.04) 14:07:42 #link https://lists.openstack.org/pipermail/openstack-discuss/2022-November/031205.html 14:07:51 #link https://etherpad.opendev.org/p/migrate-to-jammy 14:08:42 gmann, is planning to merge the migration patches on 18th November, not sure how we will be able to merge the tempest patch given the ceph job failure 14:08:51 we've 2 patches proposed, one for cinder and one for os-brick 14:09:16 the cinder change is failing with the same reason tempest is failing, some issue in the ceph job for packages 14:09:39 #link https://review.opendev.org/c/openstack/cinder/+/861609 14:09:47 #link https://review.opendev.org/c/openstack/os-brick/+/862183 14:10:03 rosmaita, proposed the os-brick patch which is failing on py39 for some test which looks random 14:10:07 probably os-brick is fine 14:10:20 we've a non-voting ceph job which is failing 14:10:35 os-brick-src-devstack-plugin-ceph-nv 14:11:15 i'll go back and look at the os-brick patch later after the new logs are available 14:11:46 cool 14:11:51 i see they've a patch up for ceph issue 14:11:53 #link https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/863524 14:12:01 Just to add a comment: devstack + ceph + Jammy don't work either (as fas as I could tell) 14:12:04 i figure the ceph job will start working once that merges 14:12:31 manually *, well.. i guess it's the same 14:13:44 so we look good there, let's wait for gate results 14:14:25 moving on, next, Call for presentations for 2023 OpenInfra summit is open 14:14:55 CFP is open for 2023 vancouver summit 14:15:04 #link https://lists.openstack.org/pipermail/openstack-discuss/2022-November/031201.html 14:15:26 if anyone is interested in presenting their topics, please submit your presentations 14:16:08 next, Upcoming releases 14:16:14 1) cinderclient for M-1 14:16:21 #link https://review.opendev.org/c/openstack/releases/+/864531 14:16:36 we've rosmaita's patch that fixes snapshot create operation 14:16:47 so looks like we should release 14:17:03 jbernard, and I can take a look 14:17:21 whoami-rajat: roger that 14:17:34 thanks 14:17:36 i already forgot about that patch, but it makes sense to release 14:18:12 :D 14:18:14 2) Cinderlib Zed (16 Dec) 14:18:21 #link https://review.opendev.org/c/openstack/releases/+/864504 14:18:35 this is tricky since our gate is broken for the whole cycle 14:19:10 rosmaita, i forgot what we discussed last week about it ... 14:19:12 i sort of volunteered to look at that, but haven't done it yet 14:19:25 oh ok, great 14:19:36 the plan was if it's an easy fix i'll do it, otherwise we will bug geguileo :) 14:19:41 geguileo, wasn't around last week so just for awareness ^ 14:20:04 sounds good 14:20:47 3) cinder tempest plugin: ussuri and victoria last 14:20:54 #link https://review.opendev.org/c/openstack/releases/+/864343 14:21:08 not sure why they've tagged it -last, maybe it's a tempest plugin convention 14:21:19 but looks like they're moving these branches to EM 14:21:58 will review that 14:22:09 last announcement, Bug Deputy and Stable release manager 14:22:09 i think the -last thing is to make it easy to automate finding the tag 14:22:22 ah ok 14:22:28 thanks 14:23:07 starting with bug deputy, no one reached out to me throughout the week so if enriquetaso doesn't have any issues, she can continue doing the great work! 14:23:45 I have not issues with this, anyway if someone would like to take the role i'm open as well 14:23:56 enriquetaso FTW!!! 14:24:13 enriquetaso++ 14:24:18 \o/ 14:24:38 great 14:24:40 now regarding release liaison 14:25:02 enriquetaso: Thank you! 14:25:03 jbernard, has been doing great work, he managed to get all stable branches released and also worked on getting wallaby to EM 14:25:25 jbernard++ 14:25:42 so i will leave it upto jbernard if he would like to continue the work as release liaison for this cycle as well 14:25:56 jbernard++ 14:26:17 i do, i think i finally understand all the pieces now, going forward should be much easier 14:26:32 great 14:26:51 thanks enriquetaso and jbernard for the great work!! 14:27:33 moving on to topics now 14:27:56 #topic Discrepancy in volume create arguments "name" and "size" 14:28:09 please see last 3-4 comments here 14:28:11 #link https://review.opendev.org/c/openstack/python-openstackclient/+/857670 14:28:31 currently in OSC, name is a positional argument and size (--size) is an optional one 14:28:58 Stephen changed the OSC behavior to make name accept 0 or more arguments but it's still positional arg 14:29:19 the --size argument is mandatory to pass if we don't specify create from snapshot or volume 14:29:27 so functionality wise it works fine 14:29:51 my ask was to modify the size as a positional argument since it's needed when we are creating a normal volume (not from any source) 14:30:05 and make name as optional one (--name) since name is never mandatory 14:30:25 maybe I'm stretching it but the design doesn't look convincing to me on the OSC side 14:30:32 i guess your worry is that people moving from cinderclient -> osc will wind up with a bunch of volumes named "1024" 14:30:32 wanted to know what the team thinks about it 14:31:04 rosmaita, that's a possibility 14:31:43 the transition from cinderclient -> OSC should be seamless, at least for the main commands like volume create 14:32:05 one would hope so! 14:32:48 I've a local patch which does change this behavior but no point working on it any further if it just seems an issue only to me 14:32:53 well, fwiw, i support you on this 14:33:26 it would help if the osc team were a little more attentive to our CLI to ease the transition 14:33:41 your idea makes sense to me 14:34:19 I think we discussed this many time ago and I think name shouldn't be mandatory but I need to review that patch because I haven't 14:34:31 thanks, I will complete and push that patch then 14:34:44 enriquetaso, name isn 14:34:48 sorry 14:35:19 sorry, i'll check the patch first whoami-rajat 14:35:21 enriquetaso: it was, but stephen listened to us and made it optional recently 14:35:39 so this is a slightly different issue 14:36:01 enriquetaso, name isn't mandatory but positional so you can say "openstack volume create 1234 --size 1" but it's opposite to our structure in cinderclient 14:36:49 that's all i had for this, let's move to next topic 14:36:56 #topic Outreachy Update 14:36:58 do we have any cinder member also part of the osc core (in order to vote i mean)? I don't know how much I agree with the fact that the owner also approves the patch, but I think that discussion is for another time. 14:37:25 i think the osc is a small group 14:37:28 that's me 14:37:39 Sad news: Cinder only had one potential contribution this round. The applicant has chosen another project to work on so we don't have another applicant to pick as an intern. 14:37:42 yeah, we discussed it before, let's reiterate this discussion during midcycle 14:37:52 We are out of this Outreachy round :( 14:38:16 yeah and that's really bad, she was a promising candidate 14:38:28 yes, she was really good 14:38:52 i think one of the best I can remember, really big first contribution 14:39:26 enriquetaso: thanks for trying! 14:39:31 guess, we have to continue with the functional test ourself. I'll work on that 14:39:32 yep, so we won't be having outreachy session during this quarter right? 14:39:40 nop 14:39:46 rosmaita: ++ 14:40:15 ack, thanks for guiding me through the outreachy process enriquetaso 14:40:45 applicants really struggle to deploy devstack and make something work. most drop out at that time 14:41:00 whoami-rajat, maybe we can try again next year 14:41:30 sure 14:41:36 that's all for me 14:41:55 thanks again 14:42:12 that's all the topics we had for today, moving to open discussion 14:42:16 #topic open discussion 14:42:27 Execuse me, I'd like to bring attention in this patch here. I made a request in the reply to the comment, and I want to know whether it is feasible. 14:42:37 https://review.opendev.org/c/openstack/cinder/+/847730 14:43:00 Since the difference between current community and our lateset version is large, and I'd like to submit and merge them as soon as possible, so could you please review and comment the patches at your earliest convenience? 14:43:35 i see hemna has a -1 there 14:44:21 inori: are you saying that what you want to do is submit a series of patches, and after that, bump the version? because otherwise the "community" and "enterprise" drivers will have different versions? 14:44:37 (saying in your response to walt on the patch, i mean) 14:44:42 (o/ for later open-discussion quick topic) 14:44:52 Yes, he said that we should add versioning information. But the current community code is old, which is quite different from our latest version 14:45:58 ok, and you want to add the features to the community driver in a different order than the "enterprise" version numbers 14:46:32 Yes, does it is feasible? 14:46:44 i see another change for qos support: https://review.opendev.org/c/openstack/cinder/+/703770 14:48:17 inori: the current community driver is 1.3.0 .... what will the version be after you make all the changes? 14:48:32 703770‘s time is a little long, so I submitted 847730 again 14:49:03 it will be 1.7 14:49:46 inori, it is always good to continue in the same patch to preserve history but if you plan to move forward with 847730, would be good to abandon 703770 to avoid confusion 14:50:01 well, how about doing 1.3.1, 1.3.2, 1.3.3, etc, and then bump them both to 2.0 when you have reached functional equivalence? 14:50:56 I see. I'll try to abandon 703770 later. 14:52:15 but some function has been changed over and over again in the history versions. 14:52:19 in any case, i agree with hemna that we want some kind of version history in the file about the features that are being added 14:54:13 In other words, the question I made in 847730 may not be very feasible? 14:55:38 rosmaita, and hemna seems to disagree with the approach so probably you will need to change it 14:56:08 inori: let me try to say this another way 14:56:20 you already have 2 different drivers 14:57:10 so the version history is going to be different 14:57:37 i mean, is the plan that you will get all the code from the "enterprise" driver merged upstream, and then you will only have one driver? 14:58:58 We have FC and iSCSI driver, and the version of them are synchronized. 14:59:59 The community is 1.3.0, and now our latest is 1.7, and many changes have taken place in the process. 15:00:00 well, that's ok, it looks like you keep the version number in the common file 15:01:12 so it may be a little difficult to submit them in the order of versions. 15:01:16 yeah, but inori, we are interested in knowing what version of the upstream driver is being used ... i don't think we can wait until everything has merged, because all committed changes will be released when Antelope is released 15:02:08 inori: you don't have to submit them in the order of versions ... you are just submitting a series of patches. Just bump the openstack driver from 1.3.0 to 1.3.0.1, and then to 1.3.0.2, etc 15:02:24 and when you reach equivalence, you can bump it to whatever you want 15:02:41 and we're out of time, would be good to continue the discussion in cinder channel (after the BS meeting) or on the patch 15:02:49 thanks everyone for joining 15:02:54 #endmeeting