14:00:41 <rosmaita> #startmeeting cinder 14:00:42 <openstack> Meeting started Wed Jul 22 14:00:41 2020 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:45 <openstack> The meeting name has been set to 'cinder' 14:00:57 <rosmaita> #topic roll call 14:01:03 <e0ne> hi 14:01:05 <rajinir> hi 14:01:11 <LiangFang> hi 14:01:14 <whoami-rajat__> Hi 14:01:14 <TusharTgite> hi 14:01:17 <jungleboyj> o/ 14:01:21 <geguileo> hi! 14:01:25 <tosky> hi 14:01:37 <rosmaita> good turnout! it's been so quiet in the cinder channel, i thought maybe i had the wrong day 14:01:46 <rosmaita> #link https://etherpad.openstack.org/p/cinder-victoria-meetings 14:01:52 <lseki> hi 14:01:53 <enriquetaso> hi 14:02:16 <rosmaita> hello everyone 14:02:26 <rosmaita> #topic announcements 14:02:47 <rosmaita> ok, next week will be our first video + IRC meeting 14:02:59 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-July/015879.html 14:03:05 <smcginnis> o/ 14:03:16 <LiangFang> great 14:03:36 <rosmaita> that email has "Video Meeting Rules", please review them 14:03:40 <enriquetaso> +1 14:03:45 <rosmaita> connection info is on the etherpad 14:04:10 <rosmaita> we'll see how it goes 14:04:17 <jungleboyj> Sounds good. 14:04:26 <rosmaita> next announcement: PyCharm licenses for openstack contributors 14:04:34 <rosmaita> http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016039.html 14:04:48 <rosmaita> i think you get a license for the "professional" edition 14:05:15 <rosmaita> Victoria Milestone-2 is next week 14:05:25 <rosmaita> new driver: https://review.opendev.org/#/c/741158/ 14:05:30 <whoami-rajat__> oh, I've been using the community version 14:05:36 <rosmaita> thanks to smcginnis for all his work reviewing that 14:05:47 <rosmaita> it looks pretty ready to go 14:05:59 <rosmaita> does anyone here have a burning need to review that driver? 14:06:08 <rosmaita> (just asking, because otherwise i will do it) 14:06:28 <rosmaita> (just don't want to duplicate effort) 14:06:30 <jungleboyj> I would like to but don't know that I will get to it. :-) 14:06:56 <rosmaita> ok, what i wanted to make sure was that someone isn't in the middle of reviewing but hasn't published comments yet 14:07:26 <rosmaita> ok, sounds like my afternoon is planned then 14:07:47 <rosmaita> final announcement: Victoria mid-cycle part 2 is week of 10 August, look for poll for meeting time on the ML next week 14:07:56 <rosmaita> actually, one more thing 14:08:19 <rosmaita> i believe that the berlin summit is going to be virtual, but i couldn't find an actual announcement on the ML 14:08:56 <e0ne> rosmaita: https://www.openstack.org/summit 14:09:02 <e0ne> Open Infrastructure Summit 14:09:02 <e0ne> Virtual Event 14:09:02 <e0ne> Oct 19-23, 2020 14:09:09 <jungleboyj> They said it during the 10th anniversary meeting. 14:09:12 <rosmaita> thanks! 14:09:32 <rosmaita> i sent an email asking the foundation to update the covid-19 page 14:09:55 <rosmaita> ok, so that's all the announcements 14:10:06 <rosmaita> #topic final Victoria specs 14:10:24 <rosmaita> first, thanks to geguileo and whoami-rajat__ for their patience on their spec 14:10:43 <rosmaita> i have been a real PITA about getting the API nailed down before implementation 14:10:58 <rosmaita> #link https://review.opendev.org/733555 14:11:12 <rosmaita> that's the project-level default volume-types spec 14:11:37 <whoami-rajat__> rosmaita: do you want me to do the NITs update after the meeting or in a followup patch? 14:11:48 <rosmaita> i think it's ready to go, but it needs another review (and of course geguileo and whoami-rajat__ can't approve it) 14:11:59 <geguileo> lol 14:12:01 <rosmaita> whoami-rajat__: i don't know that they really need to be addressed 14:12:23 <rosmaita> i just couldn't help myself 14:12:47 <rosmaita> next up: reset state robustification 14:12:57 <rosmaita> #link https://review.opendev.org/682456 14:13:17 <rosmaita> so, this is a good spec, but i don't think anyone has volunteered to pick it up 14:13:40 <rosmaita> so i propose that someone else +2 it (it has one +2) and we can merge it and then move to 'untargeted' 14:13:52 <rosmaita> in case someone comes along looking for something to do 14:13:56 <geguileo> rosmaita: I like your suggestion 14:14:21 <rosmaita> great 14:14:21 <geguileo> rosmaita: development has been picked up 14:14:27 <geguileo> rosmaita: by whoami-rajat__ 14:14:47 <geguileo> that's why he's the primary assignee ;-) 14:15:08 <whoami-rajat__> :D 14:15:40 <rosmaita> oh, i was talking about reset-state-robustification 14:15:46 <rosmaita> which i just noticed i have a -1 on 14:15:50 <geguileo> oh, my bad, sorry 14:15:52 <whoami-rajat__> are we still talking about the default types ? 14:15:54 <rosmaita> np 14:16:03 <rosmaita> sorry, i am being confusing 14:16:08 <whoami-rajat__> got it 14:16:09 <geguileo> whoami-rajat__: we are not, I'm just slow to pick up changes in the conversation XD 14:16:30 <rosmaita> my proposal is that someone other than me, geguileo, or whoami-rajat__ needs to approve https://review.opendev.org/733555 14:16:58 <rosmaita> that one already has whoami-rajat__ working on an implementation, so that is good for Victoria 14:17:12 <rosmaita> everything else will have to go into wallaby or 'untargeted' 14:17:16 <rosmaita> which reminds me 14:17:26 <rosmaita> #link https://review.opendev.org/739487 14:17:37 <rosmaita> creates the 'wallaby' directory 14:18:30 <rosmaita> ok, that's all for today about specs 14:18:58 <rosmaita> #topic finish discussion of __DEFAULT__ volume type 14:19:12 <rosmaita> #link https://etherpad.opendev.org/p/__DEFAULT___type_discussion 14:19:32 <rosmaita> i took a close look at the etherpad whoami-rajat__ was so nice as to put up 14:19:38 <rosmaita> i have a proposal for how to fix this 14:19:44 <rosmaita> and i believe it is backportable 14:20:27 <rosmaita> so please direct your attention to the ^^ etherpad 14:21:03 <rosmaita> and i will be quiet for a few minutes so people can read, particularly lines 33-56 14:21:14 <rosmaita> (though it is all excellent content) 14:26:11 * rosmaita notes that 5 minutes have passed 14:26:54 <rosmaita> any comments? 14:27:40 <jungleboyj> Everyone is stunned. 14:27:52 <rosmaita> looks like 14:28:14 <rosmaita> ok, i will paste that "what to do" into the bug, and whoami-rajat__ can fix it 14:28:32 <tosky> so basically the last volume type surviving will always be the one set in default_volume_type 14:28:39 <whoami-rajat__> sounds good to me 14:28:44 <tosky> if the two last conditions are applied together 14:29:00 <rosmaita> right 14:29:34 <rosmaita> i think this is flexible for operators and captures the original intent 14:30:24 <rosmaita> ok, we can move on 14:30:40 <rosmaita> #topic A/A support 14:30:45 <rosmaita> e0ne: that's you! 14:30:49 <e0ne> rosmaita: thanks 14:31:13 <e0ne> so, technically we've got some code in cinder to get A/A HA working 14:31:40 <e0ne> and we've got very few drivers support it: https://github.com/openstack/cinder/blob/master/doc/source/reference/support-matrix.ini#L840 14:32:00 * jungleboyj catches up ... The Default Volume Type proposal looks good to me as well. 14:32:50 <jungleboyj> Wow, that is very few. 14:32:51 <e0ne> hour code says us that A/A is not supported yet: https://review.opendev.org/#/c/741412/2/cinder/cmd/volume.py and it confuses operators 14:33:05 <e0ne> s/hour/our 14:33:44 <smcginnis> 21 14:33:55 <smcginnis> Oops. :) 14:34:11 <e0ne> so my proposal it to CI job (we discussed it during Ussuri cyccle): https://review.opendev.org/#/c/741125/ 14:34:41 <e0ne> I'm proposing this job as non-voting for now to get some time for bug fixes if any new bugs will be found 14:35:01 <e0ne> actually, I've found at least two bugs with simple tempest run in A/A mode 14:36:26 <e0ne> note: our gates are not very stable today, so some of my patches have -1 from CI after rebase 14:37:12 <rosmaita> wow, so even with no special A/A tests, you're getting failures in "normal" tempest when run A/A 14:37:23 <rosmaita> i am in favor of adding this job 14:37:28 <e0ne> all patches are in the same chain, so 'Mark Active-Active mode as supported' can be merged only the last 14:37:52 <e0ne> rosmaita: yes, I'm going to implement some real A/A related tests later this cycle 14:38:00 <rosmaita> excellent 14:38:11 <jungleboyj> That is scary. Thanks for working on that e0ne 14:38:24 <e0ne> but current tempest found that volume groups are broken in A/A mode now :( 14:39:45 <e0ne> that's all updates about A/A from my side 14:39:58 <e0ne> I'm happy to get any feedback and reviews 14:40:01 <rosmaita> ok, gives us some patches to review 14:40:08 <rosmaita> thanks e0ne, this is good work 14:40:25 <rosmaita> #topic The future of TSM backup driver 14:40:33 <rosmaita> e0ne: you have the floor 14:41:16 <e0ne> I did some refactoring yesterday and found that some TSM code is broken :( 14:41:17 <e0ne> #link https://review.opendev.org/#/c/742224/ 14:41:26 <e0ne> smcginnis suggested to remove this driver 14:42:00 <e0ne> smcginnis: I'm sorry Sean, I was AFK yesterday and read your message only today's morning 14:42:29 <e0ne> as you may know, we don't have 3rd-party CI for backup drivers 14:42:57 <e0ne> and TSM driver requires some IBM software and/or hardware 14:43:07 <e0ne> honestly, I even didn't try to run it 14:44:06 <e0ne> I think, it should be OK to deprecate it in the current cycle and remove in W 14:44:46 <rosmaita> ok, i will send something to the ML announcing our plan 14:44:58 <e0ne> rosmaita: thanks a lot! 14:44:59 <rosmaita> and i will try to figure out who the TSM maintainer is 14:45:24 <e0ne> oh.. it could be a really complicated task :) 14:45:42 <rosmaita> yeah 14:45:56 <jungleboyj> Let me know if there is anything I can do to help there. 14:45:57 <e0ne> I digged for some commit in 2013 when broken code was introduced and stopped at that point 14:46:05 <jungleboyj> I still have some contacts into IBM. 14:46:27 <jungleboyj> It is probably time to deprecate that though. 14:46:30 <e0ne> jungleboyj: would you like to setup 3rd-party CI for TSM? ;) 14:46:49 <rosmaita> we should have a discussion at the midcycle about backup driver maintenance 14:46:52 * e0ne is joking 14:46:59 <e0ne> rosmaita: +1 14:47:07 <jungleboyj> Ha ha. I am already hosting logs for 3rd Party CI runs in my basement. 14:47:07 <rosmaita> not having 3rd party CI on these is ... not great 14:47:22 <jungleboyj> Why not set up a CI? ;-) 14:47:31 <smcginnis> I'm not even sure TSM is still even a product. 14:47:42 <rosmaita> i was wondering the same thing 14:48:23 <rosmaita> anything else on this one? 14:48:54 <jungleboyj> The last release was in 2011 14:49:17 <rosmaita> ok, thanks e0ne 14:49:28 <e0ne> rosmaita: you're welcome 14:49:35 <rosmaita> #topic Last call: stable/pike 14:49:42 <jungleboyj> He he: https://en.wikipedia.org/wiki/IBM_Tivoli_Storage_Manager#/media/File:Original_IBM_Almaden_Research_WDSF_team_from_an_IBM_technical_award.jpg 14:50:07 <rosmaita> ok, there is widespread agreement that stable/ocata can go EOL and no one will miss it 14:50:20 <rosmaita> however, there was one response to my proposal to EOL stable/pike 14:50:31 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-July/015860.html 14:50:34 <e0ne> it's a party time! less code to maintain :) 14:51:12 <rosmaita> so, the situation is that there's a patch to get the stable/pike gate going again 14:51:23 <jungleboyj> smcginnis: 's favorite thing. Removing code. 14:51:40 <rosmaita> but, i don't see any point in merging that if the cinder team is not interested in maintaining stable/pike 14:51:46 <smcginnis> jungleboyj: \o/ 14:51:57 <rosmaita> i asked last week, and no one appeared interested 14:52:12 <rosmaita> so i just wanted to verify with the people present 14:52:33 <rosmaita> is anyone interested in keeping stable/pike alive? 14:52:53 <rosmaita> my feeling is that just having it open is costing me too much psychic energy 14:53:05 <tosky> just out of curiosity: is there an update on the nova side? Did they merge their part of pike patches? 14:53:09 <rosmaita> and we could better spend our time fixing A/A in current branches 14:53:16 <rosmaita> i believe they did 14:53:19 <e0ne> rosmaita: :_ 14:53:20 <tosky> (because if we drop pike, almost everyone drops pike) 14:53:21 <e0ne> :) 14:53:34 <rosmaita> but the difference is that elod volunteered to maintain nova stable/pike 14:53:44 <rosmaita> he has not volunteered to maintain cinder stable/pike 14:54:09 <rosmaita> and i don't want to keep it open unless at least one cinder core agrees to "adopt" stable/pike 14:54:46 <jungleboyj> Not it 14:55:24 <rosmaita> my opinion is that the project does not need the distraction 14:55:46 <rosmaita> and if no one even wants to review stable/pike patches, then nothing can get in there anyway 14:55:59 <jungleboyj> If eharney isn't raising a concern then I think it is safe to retire it. 14:56:12 <eharney> retiring it is good 14:56:16 <rosmaita> yay! 14:56:33 <rosmaita> ok, will send out something on the ML and propose EOL patches 14:56:50 <rosmaita> actually, in reverse order -- i will put up the patches and then send something to the ML 14:57:10 <rosmaita> ok, thanks everyone for that 14:57:14 <rosmaita> #topic open discussion 14:57:51 <Liang__> could anybody help to review the volume local cache patches:) 14:58:27 <rosmaita> i think they are very close ^^ 14:58:30 <whoami-rajat__> I've some patches which i want to request reviews for 14:58:30 <whoami-rajat__> 1) https://review.opendev.org/#/c/721206/ 14:58:30 <whoami-rajat__> 2) https://review.opendev.org/#/c/735513/ 14:58:30 <whoami-rajat__> 3) https://review.opendev.org/#/c/737035/ 14:59:44 <rosmaita> #link https://review.opendev.org/663549 14:59:48 <Liang__> rosmaita: thanks 15:00:01 <lseki> bugfix for generic NFS driver is waiting for reviews, too (they're small) https://review.opendev.org/#/q/topic:bug/1870367 15:00:03 <rosmaita> #link https://review.opendev.org/700799 15:00:25 <rosmaita> ok, please return to the meeting logs when you are looking for something to do 15:00:27 <rosmaita> #endmeeting