16:00:31 <jungleboyj> #startmeeting cinder
16:00:37 <openstack> Meeting started Wed Apr 25 16:00:31 2018 UTC and is due to finish in 60 minutes.  The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:40 <openstack> The meeting name has been set to 'cinder'
16:01:00 <jungleboyj> courtesy ping:  jungleboyj DuncanT diablo_rojo, diablo_rojo_phon, rajinir tbarron xyang xyang1 e0ne gouthamr thingee erlon tpsilva patrickeast tommylikehu eharney geguileo smcginnis lhx_ lhx__ aspiers jgriffith moshele hwalsh felipemonteiro lpetrut lseki
16:01:05 <smcginnis> 🙋
16:01:06 <Swanson> Hi
16:01:09 <gouthamr> o/
16:01:11 <xyang> hi
16:01:13 <geguileo> hi! o/
16:01:13 <tommylikehu> hi
16:01:14 <abishop> o/
16:01:14 <arnewiebalck> hi
16:01:16 <hamdyk> hi
16:01:17 <jungleboyj> smcginnis:  ?
16:01:26 <jungleboyj> @!
16:01:26 <_pewp_> jungleboyj (¬_¬)ﾉ
16:01:29 <e0ne> hi
16:01:46 <e0ne> #link https://etherpad.openstack.org/p/cinder-rocky-meeting-agendas
16:02:00 <jungleboyj> e0ne:  Thank you.
16:02:04 <e0ne> jungleboyj: just got a link by the hand
16:04:22 <jungleboyj> Ok, I think we have the crew we are going to have for today.
16:04:36 <jungleboyj> #topic announcements
16:05:02 <jungleboyj> Hopefully everyone now has heard that the next PTG will be in Denver the week of 9/10
16:05:08 <jungleboyj> <Insert train horn here>
16:05:12 <e0ne> :)
16:05:15 <tommylikehu> lol
16:05:36 <jungleboyj> I have RSVPd that Cinder will participate.  Didn't think we needed to take a vote.
16:05:52 <Swanson> Denver? Randall Flagg told me Vegas.
16:05:59 <jungleboyj> Note if you don't sign up before 5/11 the price doubles.  So, register early and often.
16:06:52 <jungleboyj> Hope to see everyone there.
16:07:10 <jungleboyj> Also, a reminder to vote for the TC if you haven't already done that.
16:07:32 <Swanson> They don't let me vote anymore. Any idea how to change that?
16:07:44 <jungleboyj> Swanson:  Push code up?
16:08:11 <jungleboyj> Or maybe you didn't see the e-mail?
16:08:27 <jungleboyj> Anyway, that is all I had for announcements.
16:08:51 <jungleboyj> #topic Rocky Review Priorities
16:09:04 <jungleboyj> I took a quick look through this this morning.
16:09:17 <smcginnis> Swanson: If you didn't vote in the last couple I think your account goes inactive.
16:09:28 <smcginnis> Swanson: IIRC, you just need to log in and click a button somewhere.
16:09:43 <jungleboyj> I have approved the Signature Verification Spec and moved it up.
16:09:59 <Swanson> smcginnis, aight. I'll look for that.
16:10:10 <jungleboyj> Otherwise it looks like there are a number of patches waiting for review.
16:10:13 <Swanson> disenfranchised
16:10:14 <jungleboyj> smcginnis:  Thanks.
16:10:56 <jungleboyj> I think that is all I have on that for this week.  Do some reviews if people have a chance.
16:11:07 <jungleboyj> geguileo:  You here?
16:11:26 <geguileo> no, I'm in hiding
16:11:38 <smcginnis> :)
16:11:40 <jungleboyj> He he.  Take it that means you don't have any updates.
16:11:43 <hamdyk> lol
16:11:45 * jungleboyj sees you
16:11:50 <geguileo> jungleboyj: correct!
16:12:02 <geguileo> jungleboyj: you just see a clone, that's not me ;-)
16:12:08 <jungleboyj> Okie Dokie.
16:12:20 <jungleboyj> geguileo:  Wow, can you teach me how to do that?  I need that good a clone.
16:12:28 <geguileo> lol
16:12:47 <tommylikehu> geguileo:  can you clone another gorka and help me to do some review?
16:13:01 <geguileo> tommylikehu: will try   XD
16:13:09 <jungleboyj> tommylikehu: ++
16:13:54 <jungleboyj> Ok, we have a lot of other topics so we will move on.
16:14:12 <jungleboyj> #topic RBD Deferred Deletion
16:14:24 <arnewiebalck> geguileo tommylikehu thx for the reviews, did anyone else have a chance to have a look?
16:14:24 <jungleboyj> arnewiebalck:
16:15:08 <jungleboyj> arnewiebalck:  You here?
16:15:13 <arnewiebalck> yep
16:15:21 <jungleboyj> Hey, the floor is your's.
16:15:23 <smcginnis> Would be good to have jbernard and eharney on there, but glad to see geguileo reviewed.
16:15:36 <eharney> i'm catching up on a lot of things this week, but i just added myself to it
16:15:41 <smcginnis> arnewiebalck: But I can take a look in a bit.
16:15:50 <arnewiebalck> smcginnis: cool
16:16:04 <geguileo> the spec looked good to me, and I think eharney liked the idea
16:16:50 <arnewiebalck> there is no rush from my side, we can give it another week and rediscuss next week when more people had a chance to have a look?
16:16:56 <smcginnis> I know other storage devices have a trash concept. I wonder if we do this if we should make sure it is not RBD specific.
16:17:08 <smcginnis> Or maybe that doesn't matter here. I need to look it over yet.
16:17:10 <jungleboyj> smcginnis:  Good thought.
16:17:18 <arnewiebalck> IIRC the trash/restore idea was not very popular
16:17:29 <smcginnis> Yeah
16:17:30 <arnewiebalck> although the trash is made for this
16:17:48 <arnewiebalck> (and we would have loved to have this this week actually :-/)
16:17:49 <smcginnis> Glad to see you went right to stein. So we do have some time to discuss this.
16:18:10 <arnewiebalck> trash/restore will require some wider changes
16:18:22 <arnewiebalck> probably a new state even
16:18:28 <smcginnis> I'm sure. Almost like a manage existing.
16:18:36 <arnewiebalck> ‘recoverable’ or sth
16:18:44 <eharney> i think there's a lot of value in implementing this just in the driver rather than changing API semantics for it
16:18:48 <jungleboyj> Yeah.  That would be a notable effort but do think it is a good idea.
16:19:01 <smcginnis> In fact, maybe we can leverage managing existing to have another "recover from trash" option.
16:19:27 <smcginnis> Yeah, I should wait until I actually understand the proposal before going off on too many tangents. I'll take a look soon/
16:19:39 <eharney> if you want to be able to recover from the trash, you're going to have to rework how encryption keys get deleted too, which will turn into a very complex security concern
16:19:41 <arnewiebalck> smcginnis: that’d be great
16:19:46 <jungleboyj> If we do it in RBD only that doesn't preclude us from expanding it.  Right?
16:20:01 <eharney> jungleboyj: right, that's my thinking
16:20:52 <arnewiebalck> I actually trimmed the whole thing down to a fast deleton as I thought “restore” would have no chance of being accepted …
16:21:12 <geguileo> arnewiebalck: that was a good idea
16:21:14 <eharney> i think that's the right approach
16:21:19 <arnewiebalck> … but, yeah, that was the original idea :)
16:21:20 <jungleboyj> arnewiebalck:  Is a smart guy.
16:21:42 <arnewiebalck> :)
16:22:16 <arnewiebalck> ok, so how do we proceed? wait for some more eyes to check?
16:22:40 <smcginnis> I think so.
16:22:44 <jungleboyj> arnewiebalck:  That sounds like the plan.  And it sounds like we are in agreement that Stein is the best target.
16:22:56 <arnewiebalck> jungleboyj: ok
16:23:24 <arnewiebalck> so we revisit next week?
16:23:34 <arnewiebalck> or is that too short?
16:23:48 <smcginnis> That works for me. I should be able to review and understand it by then.
16:23:50 <jungleboyj> I think for the time being we can work it through the spec and re-address here if things get complicated.
16:23:55 <smcginnis> ++
16:24:22 <smcginnis> Yeah, we may not need to discuss in the meeting if folks review the proposal and are fine with it.
16:24:30 <arnewiebalck> cool
16:25:28 <jungleboyj> Ok.  Anything else here?
16:26:59 <smcginnis> I think that's a no.
16:27:04 <arnewiebalck> lol
16:27:19 <jungleboyj> Ok.
16:27:30 <jungleboyj> #topic NVMEof Cinder Backend
16:27:39 <jungleboyj> hamdyk:  Your up.
16:27:44 <hamdyk> hi
16:28:10 <hamdyk> please review and merge the commits please
16:28:12 <jungleboyj> Hi.
16:28:24 <hamdyk> Mellanix CI should vote for the changes soon
16:28:36 <hamdyk> we are working with CI team to do so
16:28:50 <jungleboyj> Ok.  I haven't had a chance to look at that but we will try to get a look at it.
16:29:04 <hamdyk> we answered the commits review
16:29:25 <smcginnis> Looks like we had another new driver proposed earlier today using this as well.
16:29:32 <hamdyk> if you have questions please ask else you can move to next topic
16:29:43 <e0ne> hamdykL great! thanks for getting CI working
16:29:45 <hamdyk> yes, SPDK
16:29:51 <jungleboyj> I think we still need to see CI.
16:30:05 <jungleboyj> Once that is working we can merge it.
16:30:20 <hamdyk> I see
16:30:21 <e0ne> jungleboyj: +1
16:30:23 <hamdyk> sure that
16:31:02 <jungleboyj> So, that would be the focus now.
16:31:37 <hamdyk> yes
16:31:42 <jungleboyj> hamdyk:  Any ETA on the CI working?
16:32:04 <hamdyk> um, by tomorrow
16:32:13 <jungleboyj> Ok.  Sounds good.
16:32:50 <hamdyk> good
16:32:51 <jungleboyj> hamdyk:  Anything more on that topic?
16:33:03 <hamdyk> that's it for me
16:33:26 <jungleboyj> hamdyk:  Thank you.
16:33:43 <jungleboyj> #topic Broken Doc Build.
16:33:48 <jungleboyj> smcginnis:
16:33:55 <smcginnis> So just so everyone is aware
16:34:02 <smcginnis> Local doc builds are currently broken.
16:34:21 <smcginnis> Gate doc builds were changed, so they currently do not enable warnings-as-errors.
16:34:29 <smcginnis> So some warnings ended up getting merged.
16:34:46 <smcginnis> Which when run locally through our tox target, ends up converting those warnings to errors and failing.
16:34:55 <smcginnis> #link https://review.openstack.org/#/c/563964/
16:35:01 <smcginnis> Fixes the current warnings. ^
16:35:10 <smcginnis> And I have proposed a job change:
16:35:17 <smcginnis> #link https://review.openstack.org/#/c/564232/
16:35:28 <smcginnis> That would enable warnings-as-errors, but only for py3 runs.
16:35:52 <smcginnis> So if that is accepted, I would then put up a patch to switch our docs jobs to use that job template instead.
16:35:53 <e0ne> smcginnis: +1 to treat warnings as errors
16:36:07 <jungleboyj> smcginnis: ++
16:36:22 <smcginnis> Ideally, I hope every project eventually moves over to both running py3 and treating warnings-as-errors, but I don't think many are ready yet.
16:36:26 <jungleboyj> Thank you for fixing that.  How did the warning-as-errors in the gate get changed?
16:36:33 <smcginnis> And will probably require a fair amount of work before they can.
16:36:53 <smcginnis> The job used to try to parse out the setup.cfg file for the warning flag.
16:37:11 <smcginnis> But with the latest guidelines, the thing it was parsing should no longer be there.
16:38:00 <smcginnis> So just wanted everyone to be aware of that. And to review and approve the warning fix. ;)
16:38:25 <tbarron> smcginnis: looks good, will copy the approach
16:38:28 <jungleboyj> smcginnis:  Will look at that after the meeting.
16:38:32 <jungleboyj> tbarron:  ++
16:38:57 <smcginnis> tbarron: Cool, let me know if you run into any issues.
16:39:05 <smcginnis> jungleboyj: That's all for that topic.
16:39:15 <jungleboyj> Ok, you are on the next one too.
16:39:24 <jungleboyj> #topic Config-table directive
16:39:27 * smcginnis warms up his fingers
16:39:32 <smcginnis> OK, this one...
16:39:57 <smcginnis> Way back in the day, when we had an openstack-manuals repo, there was a command to run to generate driver config option tables.
16:40:15 <smcginnis> Since then, we have moved docs in-tree in Cinder and that tool was taken out back and shot.
16:40:36 <smcginnis> So we have a bunch of stale tables, and folks have just been manually updating them.
16:40:55 <smcginnis> To address this, I've written a sphinx extension to do what that was doing manually.
16:41:02 <smcginnis> #link https://review.openstack.org/#/q/status:open+project:openstack/cinder+branch:master+topic:config-table
16:41:36 <smcginnis> This gets rid of the static tables and just needs the directive added to the drivers docs to pull in a table with any config options in the specified modules.
16:41:44 <smcginnis> It will dynamically parse that out and generate the table.
16:42:10 <jungleboyj> Cool.
16:42:11 <e0ne> smcginnis: cool. I'll review patches later today
16:42:13 <smcginnis> It's mostly the same format, with the only changes being getting rid of the (incorrect) group line at the top and showing when things are marked as deprecated.
16:42:39 <jungleboyj> Thank you so much for working on that.
16:42:40 <smcginnis> It probably should become an oslo_config extension, but I wanted to make sure it was working for our needs first.
16:42:49 <smcginnis> Then I can look at moving it over someday if others need it.
16:42:54 <jungleboyj> smcginnis:  ++
16:43:21 <smcginnis> I have a whole series to go through yet, but I'm holding off for now to make sure everyone is OK with this approach and that we don't find any oddities.
16:43:37 <smcginnis> But I think I've covered most of the different scenarios and looks OK to me so far.
16:43:59 <jungleboyj> smcginnis:  Should we maybe send a note to the mailing list to try to get attention from other driver owners who aren't here?
16:44:05 <smcginnis> The current series starts with the addition of the directive here:
16:44:07 <smcginnis> #link https://review.openstack.org/#/c/564068/
16:44:35 <smcginnis> jungleboyj: I'm not sure if we need each driver maintainer on there since it's just a drop and replace.
16:44:46 <smcginnis> But I could if you have any concerns there.
16:45:33 <jungleboyj> If you are confident it will be ok then we don't need to.
16:45:58 <smcginnis> If anything comes up that looks concerning, let me know and I can try to pull in the driver maintainer(s).
16:46:12 <jungleboyj> Makes sense.
16:46:25 <jungleboyj> smcginnis:  Anything else on that?
16:46:36 <smcginnis> Only real difference, other than the minor table tweaks, is some of these actually add or remove config options that have changed since the last time they were manually updated.
16:46:49 <jungleboyj> smcginnis:  And that is a good thing.
16:46:55 <smcginnis> Yep!
16:47:25 <jungleboyj> Ok.  Cool.
16:47:29 <smcginnis> That's probably all on that.
16:47:44 <jungleboyj> Ok.  And finally you again:
16:47:51 <jungleboyj> topic stable/ocata broken
16:48:15 <jungleboyj> #topic stable/ocata broken
16:48:32 <jungleboyj> #link https://review.openstack.org/#/c/563628/
16:48:40 <smcginnis> Our sync up with driverfixes is blocked because of gate failures.
16:48:53 <smcginnis> So we need that ^ backport to land to fix those.
16:49:11 <smcginnis> Then we should be able to finish getting things sync in this series:
16:49:18 <smcginnis> #link https://review.openstack.org/#/q/status:open+project:openstack/cinder+branch:stable/ocata+topic:driverfixes_sync
16:49:48 <jungleboyj> smcginnis:  Done.  :-)
16:50:02 <smcginnis> I really think we need to get that moving so we get stable/ synced up with driverfixes/
16:50:16 <smcginnis> jungleboyj, e0ne: Thanks!
16:50:21 <e0ne> smcginnis: np
16:51:02 <smcginnis> So that's all, just wanted to make sure everyone was aware of what was going on there too.
16:51:02 <jungleboyj> e0ne and /or eharney  Can you guys work on approving the other patches after that moves?
16:51:14 <jungleboyj> smcginnis:  Thank you!
16:51:53 <e0ne> jungleboyj: sure, I'll help with it
16:52:07 <jungleboyj> I should do a new release after all those merge then ...
16:52:17 <e0ne> jungleboyj: +1
16:52:42 <jungleboyj> Okie dokie.
16:52:47 <jungleboyj> I think we are done then.
16:52:54 <jungleboyj> #topic Open Discussion
16:53:03 <jungleboyj> Swanson:  Look at that!  I didn't run over!
16:53:20 <jungleboyj> Anyone have anything else to discuss?
16:53:27 <Swanson> jungleboyj, Nice. Now acknowledge my Stand joke!
16:53:40 <jungleboyj> Swanson: ?
16:54:01 <smcginnis> Swanson: I thought jokes were supposed to be funny.
16:54:01 <geguileo> if anybody have time
16:54:10 <geguileo> I have a FC patch for review
16:54:12 <geguileo> https://review.openstack.org/#/c/562219/
16:54:20 <jungleboyj> smcginnis:  OOooooh Burn!
16:54:22 <Swanson> smcginnis, you say that but I think of them more as performance art.
16:54:37 <smcginnis> lol
16:54:46 <smcginnis> geguileo: Oh right, saw that come through and didn't get back to it.
16:54:54 <smcginnis> geguileo: I'll try to take a look today.
16:54:56 <e0ne> I appreciate if I can get some reviews on https://review.openstack.org/#/c/563313/. It affects cinder and nova ceph jobs
16:54:58 <geguileo> thanks!!!
16:55:26 <geguileo> e0ne's patch is a fix for a regression I introduced  :-(
16:55:52 <e0ne> geguileo: don't worry, we have a time to fix it
16:56:05 <e0ne> and backport to older releases if needed
16:56:11 <e0ne> I didn't check it yet
16:56:45 <smcginnis> e0ne, geguileo: What about Melanie's comment on there?
16:56:55 <geguileo> smcginnis: I replied in a comment
16:57:02 <geguileo> but I think that the error in the logs is valid
16:57:11 <geguileo> Because it's complaining that an image doesn't exist
16:57:11 <e0ne> geguileo's comment makes sense to me
16:57:13 <smcginnis> Ah
16:57:24 <geguileo> because we deleted it 1 second before
16:57:27 <geguileo> or something like that
16:57:37 <geguileo> I haven't looked at why the tests are doing that
16:57:41 <smcginnis> OK, probably a bad test then.
16:57:46 <geguileo> that's my guess as well
16:57:53 <jungleboyj> Yeah, I think that is what is going on there.
16:58:22 <jungleboyj> That change is fine to get in and the other issue should be in another patch if there is a problem there.
16:58:37 <jungleboyj> Anything else?
16:59:17 <jungleboyj> Ok.  Lets wrap things up on time then.
16:59:22 <jungleboyj> Thanks for a good meeting team!
16:59:31 <Swanson> toodles
16:59:31 <jungleboyj> #endmeeting