Wednesday, 2022-02-09

*** dviroel|ruck|afk is now known as dviroel|ruck00:48
*** dviroel|ruck is now known as dviroel|ruck|out00:57
*** dviroel|ruck|out is now known as dviroel|out00:57
*** hemna1 is now known as hemna07:37
*** dviroel|out is now known as dviroel|ruck11:10
*** dasm|off is now known as dasm13:17
rosmaita#startmeeting cinder14:01
opendevmeetMeeting 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
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:01
opendevmeetThe meeting name has been set to 'cinder'14:01
rosmaita#topic roll call14:01
simondodsleyo/14:01
hemnayough14:01
toskyo/14:01
jungleboyjo/14:01
fabiooliveirahi o/14:01
yuvalHey14:01
TusharTgitehi14:01
milosz_linkiewiczhi14:01
kkaras_intelhello14:01
e0nehi14:02
rosmaitagood turnout!14:02
rosmaita#link https://etherpad.opendev.org/p/cinder-yoga-meetings14:02
rosmaitalet's get started14:02
jungleboyj\o/14:02
rosmaita#topic announcements14:02
rosmaitawe discussed an operator survey (maybe at the midcycle) to get an idea of how many volumes projects tend to have14:03
rosmaita#link http://lists.openstack.org/pipermail/openstack-discuss/2022-February/027077.html14:03
rosmaitathe survey is open until 30 March14:03
rosmaitaplease spread the word to your customers14:03
rosmaitathe more data we can get, the better14:04
rosmaitanext up, it's PTL election season14:04
rosmaitagood news: we have at least one candidate!14:04
rosmaita(and it's not me)14:04
hemnaoh?14:05
geguileorosmaita: quitter14:05
jungleboyj:-)  Thank you Rajat!14:05
rosmaita:D14:05
jungleboyjgeguileo: He he he.14:05
jungleboyjrosmaita:  You lasted longer than I did.14:05
rosmaitai lasted longer than i intended to, also!14:05
simondodsleyrosmaita: what will you do with your life now???14:05
jungleboyj:-)14:06
rosmaitai will fill the void somehow14:06
* fabiooliveira :O14:06
rosmaitaok, next item14:06
rosmaitayoga os-brick release is scheduled for tomorrow14:07
rosmaitathere are still patches to be reviewed (of course)14:07
rosmaitawe will discuss more later14:07
rosmaitanext, we announced last week that we would also do stable branch releases this week14:07
rosmaitabut, i've decided to hold those to next week14:08
rosmaita#link https://etherpad.opendev.org/p/cinder-stable-releases14:08
rosmaitathe targeted stuff is on ^^14:08
rosmaitaremember, we don't allow a backport to release n unless the code is already in release n+114:08
jungleboyj++14:09
rosmaitathat is, a backport to wallaby must already have merged into xena14:09
jungleboyjrosmaita: I will take a look at that list.14:09
rosmaitajungleboyj: ty14:09
rosmaitaand finally ...14:09
rosmaitafeature freeze in 2 weeks14:09
rosmaitaso we are roughly 1 month away from the yoga coordinated release14:10
rosmaitaok, on to the regular topics14:10
rosmaitai just reordered them on the agenda14:10
rosmaita#topic third-party driver backport policy 14:10
rosmaitathis is relevant to the stable releases (obviously)14:11
rosmaitai sent out a message to the ML summarizing what we agreed on at the midcycle14:11
rosmaita#link http://lists.openstack.org/pipermail/openstack-discuss/2022-February/027021.html14:11
rosmaitai am going to paste the key point into the meeting log:14:11
rosmaitaWhen 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
rosmaita(^^ that is the key point)14:12
simondodsley++14:12
rosmaitathere are a bunch of proposed backports that don't have such a statement14:12
rosmaitasimondodsley has -1d a bunch of them14:12
rosmaitabut, even without a -1, you should not expect a backport to be reviewed/approved unless you have such a statement14:13
rosmaitaso, please take appropriate action!14:13
jungleboyjrosmaita:  Ok, so on the ones that Simon has -1d, don't merge them until they indicate they have been tested?14:13
rosmaitayes, and same thing for any others14:14
rosmaitaeven without -1s14:14
jungleboyjOk.14:14
rosmaitaunless you feel like -1'ing them 14:14
rosmaitaor -2, maybe then people will really get the message14:14
rosmaitathough i'd probably reserve -2 for inappropriate backports14:14
jungleboyjWell, that will help my review stats.  ;-)14:15
rosmaitaany questions? i don't know how else to get the word out to people14:15
rosmaitawe discussed at midcycle14:15
geguileo-1 seems like the most appropriate 14:15
rosmaitaannounced on ML14:15
jungleboyjgeguileo:  ++14:15
rosmaitamentioned it in the weekly meeting14:15
simondodsleywhen vendors see there backports not merging they will take notice...14:16
rosmaitageguileo: i agree ... we will reserve -2 for indicating a possibly inappropriate backport14:16
rosmaitasimondodsley: hope so14:16
rosmaitaok, next topic14:16
rosmaita#topic new driver review checklist update 14:16
jungleboyjrosmaita:  ++14:16
rosmaitai noticed some items missing from our review list as i was reviewing new driver patches14:17
rosmaitaso i posted an update14:17
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/82276814:17
rosmaitatake a look, and feel free to make suggestions for other things we should be looking for that aren't mentioned14:17
rosmaitathat's all for that14:18
rosmaita#topic Reset state robustification review(need a review on this series)14:18
rosmaitaTusharTgite: i think this is you14:18
rosmaitalet me paste in the links for the record14:18
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/773985/2014:18
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/804757/1114:18
TusharTgiteyes i need areview on this base patch https://review.opendev.org/c/openstack/cinder/+/773985/20 to work further for this feature14:18
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/804735/1414:19
TusharTgitethere are total 5 patch in this series so far14:20
rosmaitaquick 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 state14:21
rosmaitai didn't remember that from the spec14:22
rosmaita(https://review.opendev.org/c/openstack/cinder/+/773985/20 looks good btw)14:22
rosmaitaare you doing that in some of the other patches still?14:22
rosmaita(and i may be incorrect about the spec)14:22
jungleboyjThe 400 was handled in a separate patch that I looked at yesterday.14:23
TusharTgiterosmaita: yes we were returing 400 to reset state to same state but eharney suggest that we dont need to cover this state14:23
TusharTgitejungleboyj: yo look into snapshot api patch 14:24
rosmaitaok, cool14:24
TusharTgitevolume API is the base patch for this whole series14:24
rosmaitathanks for your patience in working on this14:24
TusharTgitethanks i want to merge this in yoga cycle so please do give me reviews on this one14:25
rosmaitavolume api patch LGTM, i will look more closely immediately after the meeting14:25
TusharTgiterosmaita: thanks14:26
rosmaitaone more question ... have we had the 400 vs. 409 (Conflict) discussion?14:26
rosmaitai seem to remember discussing this with someone, and we decided that cinder mostly uses 40014:26
TusharTgiterosmaita: no we have no discussion about that14:27
rosmaitaok, you are returning 400s, is that correct?14:27
rosmaita(i mean is that what you are doing in your patches)14:28
TusharTgiterosmaita: yes i'm returning 400 in all invalid states for reset status14:28
rosmaitaok, and actually i see this line in your spec:14:28
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
rosmaitaso that means we approved 40014:29
rosmaitagreat14:29
rosmaitajust want that to be clear so that people don't downvote the patches for not using 40914:29
rosmaita(though i may have been the only person who would do that)14:29
TusharTgiterosmaita: ok sounds good14:29
rosmaitanote to self: 400 is OK!!!14:30
rosmaitathanks TusharTgite14:30
TusharTgiterosmaita: jungleboyj thank for your review14:30
geguileorosmaita: it would be good to mention it on the commit message, and the reason why it's ok to do the change, right?14:30
rosmaitageguileo: you mean the 400?14:31
geguileorosmaita: yes, changing 409 to 40014:31
rosmaitai think it's more like changing 200 to 400, but i get your point14:31
rosmaitayeah, let's come up with a boilerplate statement TusharTgite can use on the commit messages14:32
geguileothat's even worse14:32
rosmaitayeah, but the whole point of this change is that you could "succeed" and wind up in a bad place14:32
geguileoI would assume idempotency in such a method...14:32
geguileooh, I missunderstood14:32
geguileook, so now we are just fixing the behaviour14:33
rosmaitaright, i think we have idempotency back in the latest patches14:33
rosmaitaexactly14:33
rosmaitabut, it would be good to state explicitly that we don't need a microversion bump for this14:33
geguileoand that it's failing where it was succeeding before to prevent messing things up14:34
rosmaitaright, it's the "don't allow you to shoot yourself in the foot" approach14:34
geguileolol yes14:35
rosmaitasince we have a light agenda, i would like to take 5 minutes to work this out14:35
rosmaita#link https://etherpad.opendev.org/p/no-foot-shooting14:36
TusharTgiterosmaita: ok14:36
rosmaitaok, my suggestion is up there in the etherpad14:38
rosmaitait may be overly dramatic14:38
simondodsleybring on the drama...14:40
TusharTgiterosmaita: thanks for commit msg i'll add it in existing one for better understanding14:42
rosmaitais everyone ok with the message?14:43
rosmaita#link https://etherpad.opendev.org/p/no-foot-shooting14:43
rosmaitajungleboyj: do you remember what patch it was were we had the 400/409 discussion? it was about a month ago14:43
jungleboyjA month ago?  That is like a year in my life.  ;-)14:44
jungleboyjI don't remember off the top of my head.14:44
rosmaitayeah, me too, i can barely remember last week14:44
rosmaitaok, 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 better14:45
rosmaitaok, anything else on this topic?14:46
rosmaitaTusharTgite: add that paragragraph at the bottom of your commit messages, just before the "partially-implements" tag14:47
rosmaitaok, moving on14:47
jungleboyjrosmaita:  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
TusharTgiterosmaita: ok14:47
rosmaitayes, take a note for API version 4!14:47
rosmaita#topic yoga os-brick release14:47
rosmaitahere's what's available:14:47
rosmaitahttp://tiny.cc/brick-patches14:48
rosmaitaok, looking at that, we need to get the nvmeof and lightos changes reviewed14:50
rosmaitathe NVMeOF agent hasn't been updated in a while, so not that one14:50
rosmaitai 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 there14:51
rosmaitageguileo: 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 one14:52
rosmaitaseems the author believes that the case is covered14:52
rosmaitawould be good if you could respond if you disagree14:53
rosmaitais everyone clear on the review priorities, or do we need an etherpad?14:54
rosmaitathe image encryption patch https://review.opendev.org/c/openstack/os-brick/+/709432 will be held for Z, though hopefully merging very early in Z14:55
geguileorosmaita: I'm looking at the patch, and I would have to deploy a devstack to check things out14:56
geguileoand I'm using my current devstack to test a couple of other patches...14:57
rosmaitaok, thanks ... we have a little slippage time for the release14:57
rosmaitawe can talk about that offline14:57
rosmaitai should not have said that out loud14:57
rosmaitaone other patch i noticed is that https://review.opendev.org/c/openstack/os-brick/+/810419 looks like a security issue14:58
rosmaitaso i think we should prioritize that one too14:58
rosmaitaok, so: nvmeof patches (except the agent), lightos connector patches, and 81041914:59
rosmaitaanyone here who would like to make a case for any others?14:59
rosmaitaif you have time, eharney will probably buy you a beer at the PTG if you review the mypy patches15:00
rosmaitaok, we are out of time15:00
rosmaitathanks everyone!15:00
jungleboyjThanks!15:01
fabiooliveira\o15:01
rosmaita#endmeeting15:01
opendevmeetMeeting ended Wed Feb  9 15:01:18 2022 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:01
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cinder/2022/cinder.2022-02-09-14.01.html15:01
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cinder/2022/cinder.2022-02-09-14.01.txt15:01
opendevmeetLog:            https://meetings.opendev.org/meetings/cinder/2022/cinder.2022-02-09-14.01.log.html15:01
*** dviroel|ruck is now known as dviroel|out22:33

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!