14:00:22 <rosmaita> #startmeeting cinder 14:00:23 <openstack> Meeting started Wed Sep 16 14:00:22 2020 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:24 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:26 <openstack> The meeting name has been set to 'cinder' 14:00:30 <geguileo> hi! o/ 14:00:30 <rosmaita> #topic roll call 14:00:31 <eharney> hi 14:00:32 <smcginnis> o/ 14:00:33 <tosky> o/ 14:00:34 <lseki> o/ 14:00:48 <enriquetaso> o/ 14:00:50 <e0ne> hi 14:00:55 <whoami-rajat__> Hi 14:01:07 <LiangFang> hi 14:01:16 <rosmaita> good turnout! 14:01:17 <rosmaita> #link https://etherpad.openstack.org/p/cinder-victoria-meetings 14:01:20 <jungleboyj> o/ 14:01:30 <kaisers> o/ 14:01:58 <dviroel> o/ 14:02:04 <walshh_> hi 14:02:08 <rosmaita> hello everyone 14:02:15 <rosmaita> #topic announcements 14:02:24 <rosmaita> first off, some good news 14:02:30 <enriquetaso> yay 14:02:33 <rosmaita> lseki has been added to the cinder core team 14:02:36 <lseki> 🎉 14:02:42 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-September/017323.html 14:02:45 <smcginnis> Thanks for helping lseki 14:03:01 <e0ne> lseki: welcome to the team! 14:03:09 <rosmaita> this will give us a bit more reviewing bandwidth as we approach RC time 14:03:10 <geguileo> awesome! 14:03:32 <eharney> welcome! 14:03:45 <jungleboyj> Yay! 14:04:03 <rosmaita> next up, a reminder to think about Wallaby PTG topics 14:04:03 <rosmaita> #link https://etherpad.opendev.org/p/wallaby-ptg-cinder-planning 14:04:13 <rosmaita> and don't forget to register, info is on the etherpad 14:04:44 <rosmaita> also, we will be having a combined TC/PTL election season this cycle 14:04:50 <rajinir> o/ 14:04:53 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-September/017306.html 14:05:11 <LiangFang> congratulations lseki 14:05:43 <rosmaita> if you are interested in a TC slot or being cinder PTL, you can talk to jungleboyj or smcginnis or me to get more information about what that's like 14:05:44 <dviroel> lseki: congrats! 14:05:53 <andrebeltrami> lseki: congrats!!!! 14:05:57 <jungleboyj> ++ 14:06:00 <sfernand> congrats Lucio :D 14:06:16 <e0ne> rosmaita: are you going to propose yourself as a cinder PTL? 14:06:20 <rosmaita> keep those congratulations rolling in! 14:06:43 <rosmaita> e0ne: yes, i talked to my manager and red hat is supportive of me doing it again 14:06:55 <rosmaita> but, that shouldn't prevent anyone interested from stepping up! 14:06:56 <jungleboyj> rosmaita: ++ 14:07:00 <jungleboyj> King of Storage! 14:07:12 <e0ne> rosmaita: it means, we'll be in good hands at least one cycle more! 14:07:15 <rosmaita> also, if you think you might want to be PTL in a future development cycle 14:07:29 <rosmaita> let me know, there are things you can do to get ready 14:07:33 <e0ne> (I hope this phrase in English means what I think) 14:07:41 <rosmaita> (like becoming stable release czar) 14:07:55 <rosmaita> e0ne: it's not insulting, if that's what you mean 14:07:57 <rosmaita> :) 14:08:11 <e0ne> rosmaita: thanks 14:08:12 <jungleboyj> :-) 14:08:45 <rosmaita> #topic releases 14:09:06 <rosmaita> ok, lots of news here since we have been hampered by various gate problems the past 2 weeks 14:09:10 <rosmaita> so 14:09:20 <rosmaita> victoria python-brick-cinderclient-ext 1.1.1 released last week 14:09:30 <rosmaita> more or less on time 14:09:33 <rosmaita> but 14:09:42 <rosmaita> python-cinderclient is late 14:09:59 <rosmaita> and the releases PTL will start hunting me down if we don't get it out today 14:10:04 <rosmaita> so here's what's up with that 14:10:20 <rosmaita> we need to land these patches before release: 14:10:22 <smcginnis> :) 14:10:30 <rosmaita> mv 3.62 support: https://review.opendev.org/739223 (in gate now) 14:10:40 <rosmaita> release note for victoria: https://review.opendev.org/751779 (has one +2) 14:10:49 <rosmaita> update functional test jobs to use victoria python runtimes: https://review.opendev.org/752223 (needs reviews) 14:11:10 <rosmaita> ^^ is not essential, but would have to be backported anyway, so let's do it now 14:11:32 <rosmaita> smcginnis: i have a question about one of your patches: https://review.opendev.org/#/c/734615/ 14:11:49 <rosmaita> it may have been superseded by gmann's https://review.opendev.org/#/c/743082/ 14:12:08 <rosmaita> but if you could let me know what you think about that, it would be helpful 14:13:17 <smcginnis> rosmaita: Yeah, I'll abandon that. 14:13:26 <smcginnis> Though I still don't agree with the other direction. :) 14:13:30 <rosmaita> that's all for cinderclient ... hopefully the gate will cooperate and we can get it released today 14:14:05 <rosmaita> next release will be next week, cinder RC-1 for victoria 14:14:22 <rosmaita> we have two patches that need to merge before friday (they have FFEs) 14:14:36 <rosmaita> #link https://review.opendev.org/#/c/741327/ 14:14:45 <rosmaita> #link https://review.opendev.org/#/c/737707/ 14:14:57 <whoami-rajat__> rosmaita: the release patch for client doesn't have 3.62 in features https://review.opendev.org/#/c/751779/2/releasenotes/notes/victoria-release-0d9c2b43845c3d9e.yaml ? 14:15:20 <rosmaita> whoami-rajat__: because it's in your patch 14:15:26 <rosmaita> don't worry i checked 14:15:33 <whoami-rajat__> oh ok 14:15:55 <rosmaita> it's based on your patch, so if you look at the releasenotes job in HTML, you should see them all together 14:16:18 <rosmaita> ok, so the two reviews above are TOP PRIORITY 14:16:32 <rosmaita> i think geguileo and i have https://review.opendev.org/#/c/737707/ covered 14:16:45 <rosmaita> but https://review.opendev.org/#/c/741327/ needs some more eyes 14:17:39 <rosmaita> until those get merged, other patches will not be getting much attention -- so if you want reviews, you need to give reviews 14:17:59 <rosmaita> both those patches look really good, so we should be ready for RC-1 14:18:26 <rosmaita> when RC-1 is cut, the stable branch for victoria will be created 14:18:51 <rosmaita> after that point, only release-critical bugs will be eligible for victoria (after they are merged to master) 14:19:12 <rosmaita> i have noticed a lot of driver maintainers asking for reviews in #openstack-cinder 14:19:17 <rosmaita> you are all in the same boat 14:19:33 <rosmaita> it will speed up reviews of your patches if you are reviewing other driver patches 14:19:42 <rosmaita> that is the way things work here in open source land 14:19:54 <smcginnis> +1 14:20:29 <rosmaita> ok, simultaneous with RC-1 I'd like to get some stable releases out for ussuri and train 14:20:36 <tosky> (hopefully review in a way that's it's not just "+1 with no comments if I've never reviewed before") 14:20:41 <jungleboyj> +++ 14:21:13 <rosmaita> tosky: good point, we all appreciate quality reviews ... make sure you say something 14:21:20 <sfernand> ;lk 14:21:41 <rosmaita> you don't have to say something negative, but you can look at the code coverage and mention whether it's good or bad 14:21:49 <rosmaita> there are all sorts of helpful things 14:22:09 <rosmaita> just ask yourself, what kind of helpful comments would I like to get, and then leave those 14:22:26 <rosmaita> so back to stable releases 14:22:35 <rosmaita> i've got patches up to put the most recent appropriate os-brick release into requirements and lower-constraints 14:23:01 <rosmaita> these make sure OSSN-0086 is fully addressed (it took 3 separate fixes, but i think we have it now) 14:23:13 <rosmaita> the changes to stable/train and stable/ussuri addressing the upgrade-to-ussuri bug have merged 14:23:21 <rosmaita> so that's another reason why i want to get these out 14:23:38 <rosmaita> finally, i would like to include backports of https://review.opendev.org/#/c/741498/ 14:23:47 <rosmaita> but we will discuss that later in the meeting 14:24:06 <rosmaita> ok that's all the announcements, which were kind of a lot 14:24:17 <jungleboyj> rosmaita: I think I got the OSSN related patches merged. Or at least +2d 14:24:34 <rosmaita> jungleboyj: great 14:24:40 <rosmaita> #topic eharney's mypy (type checking) work 14:24:55 <rosmaita> ok, eric has been working on introducing type checking into our codebase 14:25:00 <rosmaita> which will be very nice 14:25:16 <rosmaita> but he's in rebase hell until we merge them 14:25:23 <rosmaita> so i would like to propose this: 14:25:34 <rosmaita> we should have a cinder hackday sometime between when stable/victoria is cut and before the PTG and get these done and into Wallaby before any feature development starts 14:25:56 <rosmaita> i think if we all focus and work on them in real time, that will be a big start 14:26:03 <e0ne> rosmaita: it's a good plan to follow 14:26:12 <rosmaita> but at this point i am scared to try to merge them into victoria 14:26:23 <e0ne> rosmaita: we'll have less merge conflicts once it will me merged 14:26:31 <rosmaita> right 14:26:47 <jungleboyj> That plan makes sense. 14:26:50 <eharney> sounds good to me 14:26:57 <rosmaita> i will put together some kind of poll so we can pick a day/time 14:27:02 <rosmaita> but not until next week 14:27:16 <rosmaita> #action rosmaita poll for mypy hackday 14:27:43 <rosmaita> i hope eharney is ok with that plan (i can't remember if i asked him) 14:27:48 <eharney> yes :) 14:27:56 <rosmaita> excellent 14:28:12 <rosmaita> ok, just as a reminder, here are the main 2 patches 14:28:23 <rosmaita> #link https://review.opendev.org/#/c/733620/ 14:28:34 <rosmaita> #link https://review.opendev.org/#/c/733621/ 14:29:04 <rosmaita> and there was discussion at the virtual PTG and/or virtual midcycles, you can look in the summaries 14:29:14 <rosmaita> links are at the top of the agenda page 14:29:36 <rosmaita> just so you can refresh yourself about what this mypy effort is all about 14:29:52 <rosmaita> #topic Nimble multiattach support for rocky 14:30:01 <rosmaita> ajitha: that's you 14:30:43 <ajitha> rosmaita: yes 14:31:02 <e0ne> ajitha: are you proposing to backport this patch https://review.opendev.org/#/c/740125/ to stable/rocky? 14:31:21 <rosmaita> i think the question is can it be backported at all 14:31:31 <rosmaita> but the eventual destination will be rocky 14:31:38 <jungleboyj> rosmaita: No. 14:31:41 <ajitha> yes we have customer requirement 14:32:01 <e0ne> I thinks, it's obvious that we'll put -2 on such packports 14:32:06 <rosmaita> jungleboyj: i was thinking that too, but if you look at the patch, it seems more like a bugfix 14:32:29 <e0ne> ajitha: I understand your customer needs, but we can't backport features 14:32:30 <jungleboyj> Looking. 14:32:30 <rosmaita> the change is from a driver-specific way to say that you want multiattach to the common current cinder way to do it 14:32:32 <ajitha> Before supporting nimble multiattach from cinder, we do support the multiattach from array side using multi-initiator extra-specs, recently we have removed the nimble extra-specs and included cinder way of multiattach because of bugs in the earlier multiattach operation. 14:32:38 <ajitha> https://docs.openstack.org/cinder/rocky/configuration/block-storage/drivers/nimble-volume-driver.html 14:32:55 <smcginnis> Adding new driver functionality hasn't been considered a bugfix in the past. 14:32:59 <ajitha> So this can be considered as bug fix?? 14:33:13 <e0ne> smcginnis: +2 14:33:13 <smcginnis> Unless it was for core required (expected) functionality, which this is not. 14:34:19 <smcginnis> This would also require a volume type reconfiguration on a stable upgrade by dropping the existing vendor-specific multiattach and switching over to the standard multi-attach. 14:34:23 <jungleboyj> I feel like we have had this argument for Multi-Attach before. 14:34:31 <smcginnis> This and others like it. 14:34:53 <e0ne> "this feature doen't work because it doesn't exist" is not a bug report, it's a feature request 14:34:53 <rosmaita> ok, i was not clear on this myself, that's why i wanted to open it up to the wider team 14:35:28 <rosmaita> i believe it worked, you just had to configure it differently ... at least that's my understanding 14:36:23 <jungleboyj> So ... There is nothing that stops ajitha from proposing the backport and then making it available to their customers, but it that is the vendor's choice. We can't merge it into the stable base as a community. 14:36:52 <whoami-rajat__> i think the volume type extra specs change might cause some complications with the stable branches 14:37:18 <rosmaita> ok, it sounds like most of the stable maint cores are dead set against this proposal, and for some good reasons 14:37:38 <smcginnis> By the way, there are some legitimate issues with the patch that merged to master that I commented on yesterday. Just raising visibility that those need to be fixed. 14:37:48 <jungleboyj> Yes. Sorry. 14:38:00 <jungleboyj> smcginnis: ++ 14:38:05 <rosmaita> ajitha: see smcginnis comment above ^^ 14:38:46 <rosmaita> ajitha: it would be good to get the feature fixed in master before RC-1, so please concentrate on that 14:38:48 <ajitha> smcginnis: yes sure. will update 14:38:55 <smcginnis> Thanks! 14:39:13 <rosmaita> ok, next topic is another backport issue 14:39:27 <rosmaita> #topic backport "Modify default/delete volume type logic" to ussuri and train? 14:39:29 <ajitha> rosmaita: ok sure. Thank you 14:39:41 <rosmaita> we started to talk about this last week, but didn't come to a consensus 14:39:52 <rosmaita> #link https://review.opendev.org/#/c/741498/ 14:40:20 <rosmaita> i believe it's a non-breaking change that will make the default volume-type behavior consistent from Train through Victoria 14:40:40 <rosmaita> which would be a plus for operators and support people 14:41:14 <whoami-rajat__> just to add, if people know about the rename "hack" to delete the __DEFAULT__ type, i think this will add more safety to prevent against untyped volumes 14:41:50 <rosmaita> yeah, the cat is out of the bag on that one 14:42:55 <e0ne> I'm only concerned we we'll not break upgrades with it 14:43:28 <rosmaita> e0ne: that is a good concern 14:43:32 <jungleboyj> :-( Bigger patch than I normally like to see backported, but I do see the value in doing so. 14:44:09 <rosmaita> i think that the recent patches fixing the online migrations for Train (and also Ussuri, sort of) address this 14:44:21 <e0ne> it's definitely good to have this backported 14:45:54 <rosmaita> i just found the links to the upgrade-bug patches 14:46:04 <rosmaita> #link https://review.opendev.org/748482 14:46:14 <rosmaita> #link https://review.opendev.org/748481 14:46:28 <rosmaita> ok, so it sounds like backporting is not a bad idea 14:46:53 <rosmaita> whoami-rajat__: why don't you go ahead and propose a backport to U and if people have second thoughts they can be addressed in gerrit 14:47:18 <whoami-rajat__> rosmaita: done 14:47:25 <rosmaita> :D 14:47:27 <whoami-rajat__> #link https://review.opendev.org/#/c/752253/ 14:47:39 <rosmaita> #topic updates on the status of the last legacy job 14:47:51 <rosmaita> tosky: that's you 14:48:03 <jungleboyj> whoami-rajat__: :-) What took so long? 14:48:06 <tosky> there is just one job left, the multibackend matrix, which has proven to be a bit more tricky than usual 14:48:09 <tosky> I have a WIP patch that at least executes all the combinations, even though it fails on one job and the overall result is incorrect: https://review.opendev.org/#/c/751579/ 14:48:12 <tosky> What's missing is the reporting part, but luckily it should be possible with some ansible trickery with redefining facts. 14:48:16 <tosky> My main point here is that this should hopefully be merged before branching rc1 (it can always be backported, but...) so when the releases.git patch is created, let's try to see if we can have that in, if I manage to complete the porting (it doesn't matter whether it fails, as it's experimental and it fails right now as well). 14:48:20 <tosky> In the meantime any comment - and help on understanding the failures - is more than welcome. 14:48:25 <rosmaita> congratulations and thanks for getting this down to 1 job! 14:49:07 * whoami-rajat__ had the mouse hovered over cherry pick waiting for agreement jungleboyj 14:49:07 <jungleboyj> ++ 14:49:33 <rosmaita> tosky: i think we should add this to the priorities list for RC-1 14:50:04 <rosmaita> #link https://review.opendev.org/#/c/751579/ 14:50:16 <rosmaita> so even though this is marked WIP, it is important to the project 14:50:30 <rosmaita> so let's all pitch in and try to help tosky out here 14:51:03 <rosmaita> thanks, tosky ... anything else? 14:51:22 <tosky> that's all from me 14:51:34 <rosmaita> #topic open discussion 14:51:57 <rosmaita> i will be on a brief vacation starting at 1600 utc today 14:52:10 <jungleboyj> Cool. 14:52:27 <rosmaita> i will check in occasionally, but other than looking at the priority reviews will not do much stacking 14:52:43 <jungleboyj> Out the rest of the week? 14:52:56 <rosmaita> yes, will be back bright and early monday morning 14:53:18 <jungleboyj> Cool. Taking some time off is even more important right now. 14:53:29 <jungleboyj> #MentalHealth 14:53:37 <lseki> ++ 14:53:39 <rosmaita> yes, i am really feeling it 14:53:57 <rosmaita> :) 14:54:10 <jungleboyj> smcginnis: Will take care of everything for you. ;-) 14:54:19 <Liang__> rosmaita: have a good rest:) 14:54:28 <smcginnis> Hah 14:54:43 <rosmaita> that reminds me, i need to do the cycle highlights before i leave 14:54:48 <rosmaita> Liang__: ty 14:55:32 <lseki> tardy thanks for the congrats, folks :-D 14:55:34 <jungleboyj> rosmaita: ++ 14:55:37 <lseki> so probably I'll start bothering people with dumb questions, in order to understand some core patches 14:55:41 <lseki> please be patient 14:55:55 <jungleboyj> lseki: No worries. 14:56:07 <whoami-rajat__> lseki: i missed the initial part, congratulations :) 14:56:13 <jungleboyj> We will share the hive mind . 14:56:28 <rosmaita> yes, for anyone who came in late, lseki is the newest cinder core 14:56:33 <rosmaita> congratulations! 14:57:10 <Liang__> lseki: your English is very good 14:57:18 <whoami-rajat__> rosmaita: that reminds me, i am not the newest anymore 14:57:20 <rosmaita> for anyone interested in seeing how someone can review other people's drivers, look up some of lseki's recent reviews for some good pointers 14:57:31 <rosmaita> whoami-rajat__: you are becoming a cinder old man 14:57:41 <geguileo> lol 14:57:41 <rosmaita> and actually, you were a cinder core before me! 14:58:29 <whoami-rajat__> oh, the way I'm forgetting things, surely getting old :P 14:58:31 <jungleboyj> :-) How do you think that makes eharney feel? ;-) 14:58:34 <rosmaita> in any case, it is definitely good to have someone new aboard 14:58:47 <eharney> :) 14:59:55 <rosmaita> ok, we are out of time ... thanks everyone, and happy reviewing! 15:00:02 <rosmaita> #endmeeting