14:00:25 <rosmaita> #startmeeting cinder 14:00:26 <openstack> Meeting started Wed Feb 12 14:00:25 2020 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:28 <whoami-rajat__> Hi 14:00:29 <openstack> The meeting name has been set to 'cinder' 14:00:36 <lseki> hi 14:00:37 <rosmaita> #topic roll call 14:00:42 <geguileo> hi! o/ 14:00:51 <LiangFang> hi 14:00:51 <jungleboyj> o/ 14:02:09 <rosmaita> hello everyone 14:02:13 <rosmaita> #link https://etherpad.openstack.org/p/cinder-ussuri-meetings 14:02:24 <rosmaita> ok, let's get started 14:02:34 <rosmaita> #topic announcements 14:02:58 <rosmaita> #topic announcements - virtual midcycle -art 2 planning 14:03:25 <smcginnis> o/ 14:03:28 <whoami-rajat__> Part ? 14:03:38 <whoami-rajat__> Ok got it 14:03:45 <rosmaita> whoami-rajat: ty, yes part 2 14:03:53 <senrique_> o/ 14:03:57 <senrique_> o/ 14:03:59 <rosmaita> first, I never reported on the virtual mid-cycle part 1 follow-up poll 14:04:08 <e0ne> hi 14:04:12 <enriquetaso> o/ 14:04:19 <rosmaita> hello sofia 14:04:28 <sfernand> hi 14:04:52 <rosmaita> sorry, i am having typing probs this morning 14:05:05 <tosky> o/ 14:05:06 <rosmaita> ok, closed the poll on 5 feb 14:05:10 <rosmaita> had 9 responses 14:05:26 <rosmaita> 5 from people not there, which was good, i wanted to know why they didn't attend 14:05:39 <rosmaita> 2 had conflicts suddenly come up; 1 thought not relevant; 1 didn't know about it; 1 left 15 min early 14:05:54 <rosmaita> so, maybe one extra announcement on the ML will help that 14:06:06 <rosmaita> No suggestions for improving the format. 14:06:16 <rosmaita> 1 person mentioned that 2 hours was a good length. 14:06:28 <rosmaita> People who attended rated BlueJeans connection "pretty good to excellent." 14:06:38 <rosmaita> so i think we should stick with this for part 2 14:06:52 <rosmaita> btw, i forgot to mention that jungleboyj uploaded the recording to the cinder youtube channel 14:06:57 <jungleboyj> Yeah, I think BlueJeans works well and the resulting recordings are good. 14:07:01 <jungleboyj> :-) 14:07:03 <rosmaita> #link https://youtu.be/Dz28U1pQnqA 14:07:22 <rosmaita> ok, so part 2 has been proposed for week of 16 March 14:07:32 <rosmaita> (gives us 3 weeks before final ussuri release of client libraries) 14:07:40 <rosmaita> my question today is: does anyone here know of a major conflict that week? 14:07:40 <rosmaita> (i almost planned part 1 for the week of FOSDEM, so I wanted to check) 14:08:15 <rosmaita> well, if you think of anything, let me know 14:08:20 <jungleboyj> I am not aware of conflicts 14:08:38 <jungleboyj> Can you share the link for the voting poll? 14:08:41 <rosmaita> i'll put together a poll for day/time and send it out next week (probably after the meeting) 14:08:52 <rosmaita> jungleboyj: you are 1 step ahead of me there 14:09:05 <jungleboyj> Ah, not out yet. Sorry, misinterpreted the previous message. 14:09:19 <rosmaita> np 14:09:23 <rosmaita> #topic releases 14:09:50 <rosmaita> this is week R-minus-13 and milestone 2 14:10:00 <rosmaita> #topic releases - cinder 14:10:06 <rosmaita> we have about 260 changes since the stable/train branch 14:10:13 <rosmaita> one major feature (glance multiple store support) 14:10:20 <rosmaita> i'd like to do a milestone-2 release (even though we don't have to) 14:10:27 <rosmaita> it will be our first python3-only release (would rather not wait until milestone-3 for that) 14:10:36 <rosmaita> i took a quick look at the open patches, but didn't notice anything essential for M-2 14:10:43 <rosmaita> #link https://review.opendev.org/#/q/project:openstack/cinder+status:open+branch:master 14:10:51 <rosmaita> so unless someone can point something out, i'll put up a release patch today 14:11:03 <rosmaita> (you don't have to tell me right this minute) 14:11:29 <rosmaita> i'll put up the patch around 20:00 UTC today unless i hear otherwise 14:11:39 <rosmaita> #topic releases - python-cinderclient 14:11:47 <rosmaita> the release team is pushing for a milestone-2 release 14:11:58 <rosmaita> i think it's a good idea to get the py3-only release out even though there are not a lot of changes there 14:12:05 <rosmaita> here's what we have open right now: 14:12:13 <rosmaita> #link https://review.opendev.org/#/q/project:openstack/python-cinderclient+status:open+branch:master 14:12:47 <rosmaita> there are a few patches sitting with one +2 14:13:04 <rosmaita> so if reviewers have time, please take a look 14:13:18 <rosmaita> i'll put up the patch for that one tomorrow 14:13:39 <rosmaita> (actually, the patch is there, i will just revise it) 14:13:42 <rosmaita> #topic releases - os-brick 14:13:52 <rosmaita> there are some things we should get in (like the cryptsetup password checking) 14:13:58 <rosmaita> this one we can release next week 14:14:08 <rosmaita> #link https://review.opendev.org/#/q/project:openstack/os-brick+status:open+branch:master 14:14:28 <rosmaita> so, after you look over python-cinderclient, please direct your attention to os-brick 14:14:37 <rosmaita> #topic releases - final rocky release 14:14:53 <rosmaita> we need everything merged by 21 Feb (release must be ready for 24 Feb) 14:15:07 <rosmaita> last i looked, there's not much 14:15:25 <tosky> there may be a few pending backports iirc 14:15:48 <rosmaita> right, so after you look at os-brick, please look at anything pending for stable/rocky 14:15:49 <tosky> or at least going their way to rocky, so hurry up with backports if you also want to get them in the last rocky tarball 14:16:06 <rosmaita> yes, if you have things on the way, please keep on top of them 14:16:07 <smcginnis> Would be useful to do a sweep all the way from train backports to see if there is anything that needs to go all the way back. 14:16:24 <rosmaita> any volunteers for ^^ ? 14:16:27 <rosmaita> it's a good idea 14:16:43 <smcginnis> Not the end of the world, but last chance to do it. 14:16:52 <rosmaita> rocky is indeed a long way back 14:16:54 <jungleboyj> I went through backports the other day and didn't see too many outstanding patches but may need to get more backported. 14:16:59 <smcginnis> * last chance to do it and get in an official release 14:17:36 <smcginnis> I started working on a script to automatically find bugfixes that have not been backported. Really should find where that went. 14:17:47 <jungleboyj> :-) 14:17:50 <rosmaita> right, the branch will go into "extended maintenance" mode -- no more releases, but we can still backport changes 14:18:24 <rosmaita> anyway, if you have done an important bugfix recently, please assess it for backport-worthiness and get it proposed ASAP 14:18:44 <rosmaita> ok, final announcement and then i will shut up 14:18:58 <rosmaita> #topic volume-local-cache spec 14:19:08 <rosmaita> cross-project meeting with nova about volume-local-cache is tomorrow, 13:30-14:00 UTC 14:19:16 <rosmaita> details are here: 14:19:17 <LiangFang> thanks rosmaita 14:19:19 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-February/012457.html 14:19:26 <rosmaita> LiangFang: np 14:19:52 <LiangFang> I have replied some info in: https://etherpad.openstack.org/p/volume-local-cache 14:20:10 <rosmaita> ok, i will definitely take a look before the meeting 14:20:11 <rosmaita> sorry about the weird time thing on that poll, btw 14:21:07 <rosmaita> the reason for the middle of the night in europe time choices, though, was that only new york, mumbai, and shanghai time zones were mentioned when i asked in the meeting last week who might attend 14:21:28 <rosmaita> i was trying to be creative, but forgot that it was like 2:30 am in europe 14:21:47 <smcginnis> Not really a good way to include US and China without that happening. 14:21:47 <rosmaita> anyway, 13:30 UTC seems to be ok for most people 14:22:11 <rosmaita> yes, our geographic spread makes it really challenging to schedule stuff 14:22:16 <LiangFang> not sure if eric will join or not 14:22:36 <rosmaita> you mean eric fried? 14:22:44 <LiangFang> eric honey 14:22:47 <smcginnis> Or Harney. 14:22:52 <jungleboyj> :-) 14:22:54 <LiangFang> smcginnis: yes 14:22:56 <LiangFang> :) 14:22:58 <jungleboyj> I Like Eric Honey 14:23:14 <LiangFang> ... 14:23:31 <rosmaita> well, we'll see what we can get done 14:23:50 <rosmaita> ok, on to our regular topics, bunch of stuff on the agenda 14:24:07 <rosmaita> #topic RSD Cinder Driver move to unsupported 14:24:15 <rosmaita> IvensZambrano: that's you 14:24:21 <IvensZambrano> thanks 14:24:25 <IvensZambrano> hi everyone 14:24:34 <jungleboyj> Hey IvensZambrano 14:25:19 <IvensZambrano> Due to an org change the team I belong to is moving away from some of the activites we were supporting and performing for the project 14:25:33 <rajinir> o/ 14:25:57 <IvensZambrano> that includes the RSD cinder driver, the new proposal for the Swordfish driver and the changes proposed for os-brick 14:26:08 <IvensZambrano> for NVMeoF 14:26:32 <IvensZambrano> yes rajinir? 14:26:38 <jungleboyj> :-( 14:26:42 <smcginnis> Wasn't Mellanox also participating in some of that? 14:26:47 <jungleboyj> Ok. 14:27:25 <rosmaita> well, that's kind of a bummer 14:27:27 <IvensZambrano> i'm not sure smcginnis, we were not directly interacting with them on those three items 14:27:27 <rosmaita> sorry to hear it 14:27:37 <jungleboyj> So, IvensZambrano brought this up to me last week. 14:27:41 <IvensZambrano> it is, sorry to come with such announcements 14:27:43 <smcginnis> Yeah, that's really too bad. 14:28:03 <jungleboyj> Lenovo had been leveraging some of this work but when I raised it internally didn't get any concern. 14:28:15 <jungleboyj> Thought we should bring it up here given that it is pretty new technology. 14:28:26 <jungleboyj> Does this impact anyone else? 14:28:33 <jungleboyj> Any concern from the distros? 14:28:38 <smcginnis> Anyone at companies using this, now would be a good time to reach out internally and make sure they are aware of the impact of this. 14:28:51 <jungleboyj> smcginnis: ++ 14:29:21 <rosmaita> i guess a ML announcement would be in order, too 14:29:31 <jungleboyj> rosmaita: Probably a good idea. 14:29:36 <IvensZambrano> yes please, it will be good to know if any impacts to current usages of it will come 14:29:45 <jungleboyj> I am really sorry to see the RedFish work being stopped. 14:29:47 <IvensZambrano> as we are hosting the RSD Driver CI as well 14:29:54 <jungleboyj> and Swordfish 14:29:58 <IvensZambrano> and it will go off soon 14:30:07 <e0ne> jungleboyj: +1 14:30:14 <rosmaita> ok, thanks for keeping us posted 14:30:31 <IvensZambrano> well, there is always the option of handing over to another organisation the evolution of the Swordfish driver 14:30:42 <jungleboyj> I shared the process for unsupporting. Guess whether it gets removed in the next release depends on the outcome of those other discussions. 14:30:50 <IvensZambrano> that is a nice project and it is aligned to open standards 14:30:59 <jungleboyj> IvensZambrano: ++ 14:31:52 <jungleboyj> So, don't see anyone super concerned here. 14:31:54 <rosmaita> i guess we have until M-3 to get it sorted out 14:32:10 <smcginnis> Are there storage devices supporting Swordfish yet? 14:32:11 <rosmaita> maybe everyone is shocked? 14:32:17 <jungleboyj> :-) 14:33:11 <IvensZambrano> we were using an intermediate software layer for our development but the end goal was to extend the BMCs at some point to natively support SF 14:33:44 <rosmaita> IvensZambrano: do you want to send something to the operators list, or do you want me to? 14:34:07 <jungleboyj> We should get an e-mail out there somewhere. 14:34:45 <IvensZambrano> rosmaita: what is the usual procedure? 14:34:55 <rosmaita> not sure exactly 14:35:12 <rosmaita> tell you what, we can collaborate on an etherpad and i can send something out 14:35:24 <jungleboyj> rosmaita: ++ That sounds like a good plan. 14:35:32 <IvensZambrano> rosmaita: thanks! that will be great 14:35:35 <rosmaita> i want to make sure i get the detals right, especially about who to contact if they are interested in picking up the SF work 14:35:50 <jungleboyj> IvensZambrano: Thanks for being proactive and communicating with us. 14:36:16 <rosmaita> yes, thanks 14:36:17 <IvensZambrano> now I'm more sad to leave the project with this great support 14:36:24 <rosmaita> IvensZambrano: https://etherpad.openstack.org/p/swordfish-and-stuff 14:36:27 <rosmaita> jungleboyj: ^^ 14:36:30 <jungleboyj> rosmaita: It is working ;-) 14:36:54 <IvensZambrano> rosmaita: thanks! 14:37:04 <jungleboyj> IvensZambrano: Sad to see you go. Enjoyed the bit we got to work together. 14:37:14 <IvensZambrano> jungleboyj: thanks as well for the collaboration the past few days! 14:37:32 <IvensZambrano> davidsha: anything to add? 14:38:18 <IvensZambrano> you were actively collaborating with the project from our team 14:38:56 <davidsha> It's been a pleasure working with you all, and hope we get to collaborate again in the future! 14:39:07 <rosmaita> davidsha: likewise! 14:39:23 <IvensZambrano> thank you all! 14:39:33 <rosmaita> ok, thanks for that sad announcement 14:39:38 <rosmaita> #topic Backport Policy: Should we backport the Incremental Backup patch? 14:39:51 <rosmaita> #link https://review.opendev.org/#/c/705553/ 14:39:58 <rosmaita> enriquetaso: this one is you 14:40:02 <enriquetaso> hi 14:40:08 <rosmaita> hello 14:40:14 <enriquetaso> I think there is some misunderstood between the incremental patch. The patch is changing object version and because of that It may be seen as a new feature. I would like to discuss if backporting the patch is a good idea for us : https://review.opendev.org/#/c/705553/ 14:40:25 <smcginnis> I would say that's a bug since the end user could specify --incremental and have no idea it didn't do what they expected. 14:40:41 <rosmaita> smcginnis: ++ 14:40:43 <enriquetaso> ++ 14:40:49 <enriquetaso> On the other hand, rosmaita showed me this reported bug https://bugs.launchpad.net/cinder/+bug/1862635 related to the incremental patch. I didn't tought about it at the time and having CI with c-bak disable didn't help.. So, i'm working on a fix. 14:40:50 <openstack> Launchpad bug 1862635 in Cinder "Cinder backup export broken" [Undecided,New] - Assigned to Sofia Enriquez (lsofia-enriquez) 14:40:50 <e0ne> smcginnis: +1 14:41:03 <e0ne> is't somewhere between a feature and a bug 14:41:10 <jungleboyj> smcginnis: ++ 14:41:34 <smcginnis> That does raise the point again - in most tempest runs, cinder backup has been disabled because it was constantly failing. 14:41:34 <rosmaita> yeah, my point about the bug is that we'll want to backport that too if we decide to backport the original change 14:41:39 <smcginnis> We should really fix that (tm). 14:41:39 <e0ne> from users' perspective it's a bug fix without any new CLI flags, config options or API changes 14:42:02 <jungleboyj> Should we add that to our Logo? 14:42:07 <smcginnis> ;) 14:42:09 <e0ne> :) 14:42:26 <enriquetaso> :P 14:42:28 <smcginnis> That's what the horse is looking over its shoulder at! 14:42:39 <jungleboyj> Yep. That was what I was thinking. 14:42:54 <enriquetaso> rosmaita++ sounds like a good plan 14:42:56 <rosmaita> i think e0ne clearly stated the reason why its backportable 14:43:05 <e0ne> rosmaita: thanks :) 14:43:16 <enriquetaso> e0ne++ 14:43:24 <rosmaita> ok, good ... wanted to discuss this because of an objection on enriquetaso's backport patch 14:43:35 <rosmaita> but it sounds like we are in agreement here 14:43:36 <smcginnis> One nit-ish thing - best not to propose backports to older branches until the patch is merged to the more recent branch. 14:43:57 <enriquetaso> smcginnis, yes, sorry 14:44:04 <smcginnis> I know folks want to make sure it's everywhere, but that puts the burden on reviewers to make sure things don't get messed up for stable policy. 14:44:08 <jungleboyj> smcginnis: ++ 14:44:15 <smcginnis> Not a huge deal, just wanted to point that out. 14:44:18 <jungleboyj> Otherwise it sounds fine to backport. 14:44:35 <jungleboyj> Hey, Red Hat made a logo for cinderlib. :-) 14:44:40 <rosmaita> smcginnis: thank you for noticing that 14:45:02 <enriquetaso> where is the cinderlib logo ? :O 14:45:47 <jungleboyj> In a presentation I am watching. 14:46:06 <rosmaita> ok, so the consensus is: is backportable along with the related bugfix (after the bug is fixed); needs to go in the regular backport order T->S->R 14:46:15 <rosmaita> jungleboyj: is multitasking! 14:46:17 <smcginnis> ++ 14:46:23 <jungleboyj> ++ 14:46:40 <rosmaita> ok, enriquetaso -- all your questions answered? 14:46:54 <enriquetaso> yes 14:46:56 <enriquetaso> thanks 14:47:13 <rosmaita> great, this is exciting, we can have open discussion! 14:47:18 <rosmaita> #topic open discussion 14:47:27 <rosmaita> actually, I have something 14:47:34 <whoami-rajat> i forgot to mention before, i've a patch that i want it to go to rocky release https://review.opendev.org/#/c/707343/ 14:48:37 <rosmaita> whoami-rajat: did you figure out the problem with lower-constraints ? 14:48:47 <whoami-rajat> yep, unrelated 14:48:54 <whoami-rajat> its currently passing on zuul jobs 14:49:05 <rosmaita> ok, cool -- i dind't notice the recheck 14:49:05 <jungleboyj> Ok. LGTM when it passes the check. 14:49:19 <whoami-rajat> thanks!! 14:49:40 <rosmaita> i remember looking at the patch to master, i think this should be fine 14:49:46 <jungleboyj> rosmaita: Don't need any discussion on the Datera driver? 14:49:59 <rosmaita> i haven't heard anything back yet 14:50:05 <jungleboyj> Ok. 14:50:26 <rosmaita> but not to keep people in suspense 14:50:33 <jungleboyj> :-) 14:50:50 <rosmaita> Datera has got their CI running again, and are proposing that their driver be marked supported in U 14:51:03 <rosmaita> there is some discussion on that patch 14:51:35 <rosmaita> which i can't find the link to atm 14:52:09 <rosmaita> anyway, things are kind of up in the air, but will hopefully be resolved ok 14:52:18 <smcginnis> https://review.opendev.org/#/c/704153/ 14:52:24 <rosmaita> ty 14:52:36 <rosmaita> part of the problem is python2 compatibility 14:53:24 <smcginnis> I really would like to see that broken out into multiple patches, but I think I've said enough on that patch. :| 14:53:48 <rosmaita> yes, and i did mention that to them 14:54:08 <jungleboyj> :-) 14:54:45 <rosmaita> i think part of the problem is that they'd like to fix the driver in U and then be able to have people use it with N through T also 14:54:53 <rosmaita> not sure how viable a plan that is 14:55:09 <smcginnis> Another scary part of all that. 14:55:18 <jungleboyj> :-( That is scary. 14:55:18 <rosmaita> and i think that's part of the big monolithic patch 14:55:30 <e0ne> rosmaita: +1 14:55:44 <rosmaita> because i think they're thinking they need to backport all that stuff everywhere 14:55:57 <jungleboyj> Ouch. That won't work. 14:55:59 <rosmaita> instead of having multiple small patches and only backporting the appropriate stuff 14:56:05 <jungleboyj> rosmaita: ++ 14:56:21 <jungleboyj> Maybe communicating that better to them will help. 14:56:28 <rosmaita> but, of course, it's easy to say after they've done a bunch of work: break it into smaller pieces! 14:56:39 <rosmaita> it's pretty hard to hear that when you've done all the work! 14:56:45 <jungleboyj> True. 14:56:47 <rosmaita> so we'll see what happens 14:57:10 <jungleboyj> rosmaita: Thanks for following up. 14:57:15 <jungleboyj> One other topic. 14:57:19 <smcginnis> Well, all of those changes built up over time in their GitHub fork, so theoretically they could port over commits from there piecemeal. 14:57:32 <rosmaita> ok, one thing from me for people running Train or who know such people 14:57:40 <jungleboyj> smcginnis: ++ 14:57:44 <rosmaita> this patch has been rejected for nova (i'm working on a more robust fix) 14:57:44 <rosmaita> but the release note is worth looking at: 14:57:44 <rosmaita> #link https://review.opendev.org/#/c/706298/1 14:57:55 <rosmaita> it's about a problem that arises in a really weird situation with our automatic encryption key handling (it was actually a change in glance, but really for us) 14:57:55 <rosmaita> anyway, it can be addressed immediately by a nova config change 14:57:55 <rosmaita> i will be sending out something to the operators list 14:57:55 <rosmaita> but if you are running Train, or know people who are, you might want to bring it to the appropriate person's attention 14:58:15 <rosmaita> that's it from me 14:58:20 <LiangFang> previously i mentioned there's a white paper of volume local cache. Now it is approved to publish. But it is in chinese, co-authored by China Unicom. Mainly shows the performance gain with Optane SSD. Will upload the writepaper to some website later and let team know. Does openstack.org has some place to containing write paper? 14:58:23 <rosmaita> 1.75 minutes left 14:58:27 <trident> One other topic as well if I may, I filed a bug a souple of days ago. I intend to write a patch for it but would at some point apreciate some discussions on the best way to architect the fix. - https://bugs.launchpad.net/cinder/+bug/1862636 14:58:28 <openstack> Launchpad bug 1862636 in Cinder "cinder retype corrupts volume in some scenarios" [Undecided,New] 14:58:53 <rosmaita> that doesn't sound good 14:59:00 <rosmaita> trident: thanks for reporting it 14:59:14 <rosmaita> and glad to hear you want to fix it, too! 14:59:47 <jungleboyj> rosmaita: Have you polled the team to see who is planning to be in Vancouver? 15:00:07 <jungleboyj> Think we need to let OSF know if we are going to be there soon. 15:00:13 <rosmaita> i have not, and the table planning is due feb 21 i think 15:00:28 <rosmaita> show of hands: who is probably going to be in vancouver? 15:00:36 <rosmaita> o/ 15:00:36 <jungleboyj> o/ Planning to be. 15:01:04 <enriquetaso> i don't know 15:01:18 <rosmaita> actually, i don't know either 15:01:22 <smcginnis> Probably. No official word yet. 15:01:24 <enriquetaso> u.u 15:01:27 <rosmaita> it's kind of early 15:01:27 <whoami-rajat> :P 15:01:41 <jungleboyj> Don't need to know for sure. Just need to know how many want to attend. 15:01:43 <rosmaita> i think the same size table we had in shanghai will be good 15:01:49 <jungleboyj> rosmaita: ++ 15:01:58 <rosmaita> which was what, 12 seats? 15:01:59 <whoami-rajat> rosmaita, ++ 15:02:02 <jungleboyj> Lets do a ML post and see what kind of response we get. 15:02:13 <rosmaita> ok, i'll get one out this afternoon 15:02:16 <rosmaita> thanks everyone! 15:02:19 <jungleboyj> Thanks! 15:02:22 <rosmaita> jungleboyj: thanks for the reminder 15:02:27 <jungleboyj> :-) 15:02:27 <whoami-rajat> Thanks! 15:02:31 <jungleboyj> That is what I am here for. 15:02:36 <rosmaita> :D 15:02:39 <rosmaita> #endmeeting