14:01:14 <rosmaita> #startmeeting cinder 14:01:14 <opendevmeet> Meeting started Wed Feb 9 14:01:14 2022 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:14 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:14 <opendevmeet> The meeting name has been set to 'cinder' 14:01:20 <rosmaita> #topic roll call 14:01:24 <simondodsley> o/ 14:01:26 <hemna> yough 14:01:26 <tosky> o/ 14:01:26 <jungleboyj> o/ 14:01:27 <fabiooliveira> hi o/ 14:01:27 <yuval> Hey 14:01:32 <TusharTgite> hi 14:01:34 <milosz_linkiewicz> hi 14:01:39 <kkaras_intel> hello 14:02:11 <e0ne> hi 14:02:34 <rosmaita> good turnout! 14:02:43 <rosmaita> #link https://etherpad.opendev.org/p/cinder-yoga-meetings 14:02:48 <rosmaita> let's get started 14:02:52 <jungleboyj> \o/ 14:02:55 <rosmaita> #topic announcements 14:03:24 <rosmaita> we discussed an operator survey (maybe at the midcycle) to get an idea of how many volumes projects tend to have 14:03:37 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2022-February/027077.html 14:03:44 <rosmaita> the survey is open until 30 March 14:03:58 <rosmaita> please spread the word to your customers 14:04:14 <rosmaita> the more data we can get, the better 14:04:41 <rosmaita> next up, it's PTL election season 14:04:51 <rosmaita> good news: we have at least one candidate! 14:04:54 <rosmaita> (and it's not me) 14:05:08 <hemna> oh? 14:05:17 <geguileo> rosmaita: quitter 14:05:22 <jungleboyj> :-) Thank you Rajat! 14:05:23 <rosmaita> :D 14:05:29 <jungleboyj> geguileo: He he he. 14:05:40 <jungleboyj> rosmaita: You lasted longer than I did. 14:05:51 <rosmaita> i lasted longer than i intended to, also! 14:05:57 <simondodsley> rosmaita: what will you do with your life now??? 14:06:02 <jungleboyj> :-) 14:06:18 <rosmaita> i will fill the void somehow 14:06:31 * fabiooliveira :O 14:06:58 <rosmaita> ok, next item 14:07:10 <rosmaita> yoga os-brick release is scheduled for tomorrow 14:07:20 <rosmaita> there are still patches to be reviewed (of course) 14:07:25 <rosmaita> we will discuss more later 14:07:49 <rosmaita> next, we announced last week that we would also do stable branch releases this week 14:08:04 <rosmaita> but, i've decided to hold those to next week 14:08:14 <rosmaita> #link https://etherpad.opendev.org/p/cinder-stable-releases 14:08:24 <rosmaita> the targeted stuff is on ^^ 14:08:55 <rosmaita> remember, we don't allow a backport to release n unless the code is already in release n+1 14:09:08 <jungleboyj> ++ 14:09:16 <rosmaita> that is, a backport to wallaby must already have merged into xena 14:09:22 <jungleboyj> rosmaita: I will take a look at that list. 14:09:29 <rosmaita> jungleboyj: ty 14:09:35 <rosmaita> and finally ... 14:09:46 <rosmaita> feature freeze in 2 weeks 14:10:01 <rosmaita> so we are roughly 1 month away from the yoga coordinated release 14:10:18 <rosmaita> ok, on to the regular topics 14:10:43 <rosmaita> i just reordered them on the agenda 14:10:54 <rosmaita> #topic third-party driver backport policy 14:11:07 <rosmaita> this is relevant to the stable releases (obviously) 14:11:24 <rosmaita> i sent out a message to the ML summarizing what we agreed on at the midcycle 14:11:32 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2022-February/027021.html 14:11:54 <rosmaita> i am going to paste the key point into the meeting log: 14:12:05 <rosmaita> When a vendor driver patch backport is proposed, we would like to see a clear statement on the gerrit review that the patch has been tested in an appropriate environment. 14:12:15 <rosmaita> (^^ that is the key point) 14:12:20 <simondodsley> ++ 14:12:47 <rosmaita> there are a bunch of proposed backports that don't have such a statement 14:12:55 <rosmaita> simondodsley has -1d a bunch of them 14:13:18 <rosmaita> but, even without a -1, you should not expect a backport to be reviewed/approved unless you have such a statement 14:13:27 <rosmaita> so, please take appropriate action! 14:13:50 <jungleboyj> rosmaita: Ok, so on the ones that Simon has -1d, don't merge them until they indicate they have been tested? 14:14:01 <rosmaita> yes, and same thing for any others 14:14:07 <rosmaita> even without -1s 14:14:12 <jungleboyj> Ok. 14:14:15 <rosmaita> unless you feel like -1'ing them 14:14:32 <rosmaita> or -2, maybe then people will really get the message 14:14:43 <rosmaita> though i'd probably reserve -2 for inappropriate backports 14:15:00 <jungleboyj> Well, that will help my review stats. ;-) 14:15:05 <rosmaita> any questions? i don't know how else to get the word out to people 14:15:11 <rosmaita> we discussed at midcycle 14:15:15 <geguileo> -1 seems like the most appropriate 14:15:16 <rosmaita> announced on ML 14:15:21 <jungleboyj> geguileo: ++ 14:15:24 <rosmaita> mentioned it in the weekly meeting 14:16:00 <simondodsley> when vendors see there backports not merging they will take notice... 14:16:03 <rosmaita> geguileo: i agree ... we will reserve -2 for indicating a possibly inappropriate backport 14:16:24 <rosmaita> simondodsley: hope so 14:16:32 <rosmaita> ok, next topic 14:16:43 <rosmaita> #topic new driver review checklist update 14:16:46 <jungleboyj> rosmaita: ++ 14:17:01 <rosmaita> i noticed some items missing from our review list as i was reviewing new driver patches 14:17:07 <rosmaita> so i posted an update 14:17:16 <rosmaita> #link https://review.opendev.org/c/openstack/cinder/+/822768 14:17:37 <rosmaita> take a look, and feel free to make suggestions for other things we should be looking for that aren't mentioned 14:18:06 <rosmaita> that's all for that 14:18:25 <rosmaita> #topic Reset state robustification review(need a review on this series) 14:18:33 <rosmaita> TusharTgite: i think this is you 14:18:39 <rosmaita> let me paste in the links for the record 14:18:48 <rosmaita> #link https://review.opendev.org/c/openstack/cinder/+/773985/20 14:18:56 <rosmaita> #link https://review.opendev.org/c/openstack/cinder/+/804757/11 14:18:59 <TusharTgite> yes i need areview on this base patch https://review.opendev.org/c/openstack/cinder/+/773985/20 to work further for this feature 14:19:05 <rosmaita> #link https://review.opendev.org/c/openstack/cinder/+/804735/14 14:20:09 <TusharTgite> there are total 5 patch in this series so far 14:21:46 <rosmaita> quick question ... this doesn't happen on https://review.opendev.org/c/openstack/cinder/+/773985/20 (i just looked), but at some point you were returning 400 for resetting state to the current state 14:22:00 <rosmaita> i didn't remember that from the spec 14:22:09 <rosmaita> (https://review.opendev.org/c/openstack/cinder/+/773985/20 looks good btw) 14:22:37 <rosmaita> are you doing that in some of the other patches still? 14:22:47 <rosmaita> (and i may be incorrect about the spec) 14:23:39 <jungleboyj> The 400 was handled in a separate patch that I looked at yesterday. 14:23:46 <TusharTgite> rosmaita: yes we were returing 400 to reset state to same state but eharney suggest that we dont need to cover this state 14:24:16 <TusharTgite> jungleboyj: yo look into snapshot api patch 14:24:36 <rosmaita> ok, cool 14:24:38 <TusharTgite> volume API is the base patch for this whole series 14:24:53 <rosmaita> thanks for your patience in working on this 14:25:26 <TusharTgite> thanks i want to merge this in yoga cycle so please do give me reviews on this one 14:25:55 <rosmaita> volume api patch LGTM, i will look more closely immediately after the meeting 14:26:10 <TusharTgite> rosmaita: thanks 14:26:22 <rosmaita> one more question ... have we had the 400 vs. 409 (Conflict) discussion? 14:26:55 <rosmaita> i seem to remember discussing this with someone, and we decided that cinder mostly uses 400 14:27:04 <TusharTgite> rosmaita: no we have no discussion about that 14:27:52 <rosmaita> ok, you are returning 400s, is that correct? 14:28:00 <rosmaita> (i mean is that what you are doing in your patches) 14:28:39 <TusharTgite> rosmaita: yes i'm returning 400 in all invalid states for reset status 14:28:50 <rosmaita> ok, and actually i see this line in your spec: 14:29:01 <rosmaita> "os-reset-status actions on volumes, snaps, backups, groups, will now return a 400 in some cases where they would previously succeeded. This does not require a microversion bump." 14:29:09 <rosmaita> so that means we approved 400 14:29:13 <rosmaita> great 14:29:34 <rosmaita> just want that to be clear so that people don't downvote the patches for not using 409 14:29:47 <rosmaita> (though i may have been the only person who would do that) 14:29:48 <TusharTgite> rosmaita: ok sounds good 14:30:01 <rosmaita> note to self: 400 is OK!!! 14:30:12 <rosmaita> thanks TusharTgite 14:30:39 <TusharTgite> rosmaita: jungleboyj thank for your review 14:30:44 <geguileo> rosmaita: it would be good to mention it on the commit message, and the reason why it's ok to do the change, right? 14:31:00 <rosmaita> geguileo: you mean the 400? 14:31:16 <geguileo> rosmaita: yes, changing 409 to 400 14:31:34 <rosmaita> i think it's more like changing 200 to 400, but i get your point 14:32:04 <rosmaita> yeah, let's come up with a boilerplate statement TusharTgite can use on the commit messages 14:32:05 <geguileo> that's even worse 14:32:32 <rosmaita> yeah, but the whole point of this change is that you could "succeed" and wind up in a bad place 14:32:38 <geguileo> I would assume idempotency in such a method... 14:32:53 <geguileo> oh, I missunderstood 14:33:11 <geguileo> ok, so now we are just fixing the behaviour 14:33:12 <rosmaita> right, i think we have idempotency back in the latest patches 14:33:15 <rosmaita> exactly 14:33:39 <rosmaita> but, it would be good to state explicitly that we don't need a microversion bump for this 14:34:05 <geguileo> and that it's failing where it was succeeding before to prevent messing things up 14:34:39 <rosmaita> right, it's the "don't allow you to shoot yourself in the foot" approach 14:35:06 <geguileo> lol yes 14:35:57 <rosmaita> since we have a light agenda, i would like to take 5 minutes to work this out 14:36:00 <rosmaita> #link https://etherpad.opendev.org/p/no-foot-shooting 14:36:31 <TusharTgite> rosmaita: ok 14:38:49 <rosmaita> ok, my suggestion is up there in the etherpad 14:38:54 <rosmaita> it may be overly dramatic 14:40:16 <simondodsley> bring on the drama... 14:42:58 <TusharTgite> rosmaita: thanks for commit msg i'll add it in existing one for better understanding 14:43:12 <rosmaita> is everyone ok with the message? 14:43:21 <rosmaita> #link https://etherpad.opendev.org/p/no-foot-shooting 14:43:53 <rosmaita> jungleboyj: do you remember what patch it was were we had the 400/409 discussion? it was about a month ago 14:44:21 <jungleboyj> A month ago? That is like a year in my life. ;-) 14:44:31 <jungleboyj> I don't remember off the top of my head. 14:44:40 <rosmaita> yeah, me too, i can barely remember last week 14:45:53 <rosmaita> ok, well, people can trust us that we did discuss this about a different patch, and it turns out that the Block Storage API uses 400 even when you think 409 might be better 14:46:03 <rosmaita> ok, anything else on this topic? 14:47:09 <rosmaita> TusharTgite: add that paragragraph at the bottom of your commit messages, just before the "partially-implements" tag 14:47:13 <rosmaita> ok, moving on 14:47:19 <jungleboyj> rosmaita: Right, we had several people weigh in and decided that enforcing 409 now would be inconsistent and that changing them all would be a large undertaking. 14:47:32 <TusharTgite> rosmaita: ok 14:47:40 <rosmaita> yes, take a note for API version 4! 14:47:49 <rosmaita> #topic yoga os-brick release 14:47:58 <rosmaita> here's what's available: 14:48:12 <rosmaita> http://tiny.cc/brick-patches 14:50:09 <rosmaita> ok, looking at that, we need to get the nvmeof and lightos changes reviewed 14:50:50 <rosmaita> the NVMeOF agent hasn't been updated in a while, so not that one 14:51:37 <rosmaita> i haven't left a comment on https://review.opendev.org/c/openstack/os-brick/+/822642 , but i'm not sure about the way the retry is being handled there 14:52:32 <rosmaita> geguileo: https://review.opendev.org/c/openstack/os-brick/+/811718 has a -1 from you ... don't know if you saw the response to your comment on that one 14:52:44 <rosmaita> seems the author believes that the case is covered 14:53:18 <rosmaita> would be good if you could respond if you disagree 14:54:42 <rosmaita> is everyone clear on the review priorities, or do we need an etherpad? 14:55:40 <rosmaita> the image encryption patch https://review.opendev.org/c/openstack/os-brick/+/709432 will be held for Z, though hopefully merging very early in Z 14:56:32 <geguileo> rosmaita: I'm looking at the patch, and I would have to deploy a devstack to check things out 14:57:03 <geguileo> and I'm using my current devstack to test a couple of other patches... 14:57:04 <rosmaita> ok, thanks ... we have a little slippage time for the release 14:57:31 <rosmaita> we can talk about that offline 14:57:39 <rosmaita> i should not have said that out loud 14:58:02 <rosmaita> one other patch i noticed is that https://review.opendev.org/c/openstack/os-brick/+/810419 looks like a security issue 14:58:23 <rosmaita> so i think we should prioritize that one too 14:59:08 <rosmaita> ok, so: nvmeof patches (except the agent), lightos connector patches, and 810419 14:59:26 <rosmaita> anyone here who would like to make a case for any others? 15:00:17 <rosmaita> if you have time, eharney will probably buy you a beer at the PTG if you review the mypy patches 15:00:44 <rosmaita> ok, we are out of time 15:00:48 <rosmaita> thanks everyone! 15:01:12 <jungleboyj> Thanks! 15:01:14 <fabiooliveira> \o 15:01:18 <rosmaita> #endmeeting