17:00:05 <ildikov> #startmeeting cinder-nova-api-changes
17:00:05 <openstack> Meeting started Mon Jun 27 17:00:05 2016 UTC and is due to finish in 60 minutes.  The chair is ildikov. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:09 <openstack> The meeting name has been set to 'cinder_nova_api_changes'
17:00:18 <ildikov> scottda ildikov DuncanT ameade cFouts johnthetubaguy jaypipes takashin alaski e0ne jgriffith tbarron andrearosa hemna erlon mriedem gouthamr ebalduf patrickeast smcginnis diablo_rojo gsilvis
17:00:37 <scottda> hi
17:00:41 <DuncanT> hi
17:00:52 <ildikov> hi :)
17:01:06 <jgriffith> o/
17:01:25 <mriedem> o/
17:02:02 <ildikov> it seems hemna is on vacation, so I guess we can start
17:02:24 <ildikov> I did not really modify the agenda on the etherpad #link https://etherpad.openstack.org/p/cinder-nova-api-changes
17:02:39 <ildikov> as we still have the refactor and the testing as the two major items
17:02:50 <ildikov> jgriffith: do you have any updates?
17:03:20 <jgriffith> ildikov: just sorting through some things on the tempest code.  I'd really appreciate some feedback on the cinder side, and merge if we're moving forward
17:03:30 <jgriffith> ildikov: of course, asking for help with unit tests :)
17:03:39 <jgriffith> ildikov: and versioning mess
17:03:51 <scottda> jgriffith: O
17:03:56 <ildikov> jgriffith: unit test in Nova or Cinder or both? :)
17:03:57 <smcginnis> o/
17:04:04 <jgriffith> cinder to start
17:04:05 <scottda> I'll add some code for microversioning to your patches.
17:04:20 <scottda> Assuming that's the "versioning mess"
17:04:21 <jgriffith> ie if folks want to speed things up do follow ups
17:04:28 <jgriffith> on the Cinder side
17:04:30 <ildikov> jgriffith: ok, we can sync up after the meeting on that so we don't overwrite each other's changes
17:05:21 <ildikov> scottda: yeah, that's the one and I think you're the best to teach all of us here on how to do that properly, so thanks much in advance
17:05:41 <jgriffith> ildikov: BTW, you appear to be correct about the cross project changes
17:05:56 <jgriffith> ildikov: for some reason I thought that just *worked* now
17:05:59 <scottda> ildikov: I'll work on the Cinder side. Eventually we need to figure out the Nova code to use Cinder microversions. This will be the first time for that.
17:06:13 <ildikov> jgriffith: the client code you mean?
17:06:22 <jgriffith> ildikov: yes
17:06:45 <jgriffith> although... maybe there's something else still wrong.  Running tempest full locally again
17:07:36 <ildikov> jgriffith: Devstack did not seem to use the source for clients to install
17:07:56 <jgriffith> ildikov: right, I believe the default is pip only
17:07:57 <jgriffith> anywho...
17:07:59 <smcginnis> Do you need to use SOURCE_FROM_GIT or whatever that config is?
17:08:16 <jgriffith> smcginnis: yeah,  but I can't tell gate to do that :)
17:08:23 <jgriffith> smcginnis: that's what I do locally
17:08:34 <smcginnis> jgriffith: Yeah, that does make it a bit tricky then!
17:08:49 <jgriffith> smcginnis: just making sure I didn't miss something, but when I ran locally it all looked "ok"
17:09:07 <jgriffith> granted I only ran "volumes"
17:09:17 <ildikov> smcginnis: right, I know I saw that somewhere, thanks for reminding :)
17:09:32 <smcginnis> So we need a client released with the changes before we can actual implement the support for it.
17:09:43 <mriedem> you should be able to test the client changes + the nova changes
17:09:47 <mriedem> with some throw away patch
17:09:55 <mriedem> which would probably be a devstack change
17:10:06 <mriedem> GIT_BRANCH for cinderclient would be the ref of the cinderclient change probably
17:10:19 <mriedem> and the devstack change would depend on the nova change, which depends on the cinderclient change
17:10:27 <mriedem> the devstack DNM would actually depend on nova and cinder changes both
17:10:31 <jgriffith> mriedem: yeah, so like create a job specifically for this effort?
17:10:34 <smcginnis> mriedem: Oh, that's so crazy it might actually work. :)
17:10:39 <mriedem> jgriffith: not a job, just a throw away default config
17:10:47 <mriedem> smcginnis: we do it frequently
17:10:52 <ildikov> smcginnis: lol :)
17:10:59 <scottda> jgriffith: Just change SOURCE_FROM_GIT=blah in devstack
17:11:00 <jgriffith> mriedem: hmm... ok, I'll have to pick your brain (or someones) on how to set that up
17:11:01 <smcginnis> Makes sense.
17:11:09 <mriedem> scottda was doing something similar the other day
17:11:19 <scottda> Yeah, thanks mriedem that worked great
17:11:20 <jgriffith> mriedem: scottda ok, perfect
17:13:25 <ildikov> cool, this will make life a bit easier from early testing perspective, thanks
17:14:57 <ildikov> jgriffith and/or scottda will you set this up?
17:15:20 <jgriffith> ildikov: I'll work with scottda after the meeting if he's got time?
17:15:28 <scottda> yes, I've time
17:15:41 <ildikov> jgriffith: cool, tnx
17:16:52 <ildikov> jgriffith: do you have more work items for the Cinder side things we would need to find assignees for?
17:17:34 <jgriffith> ildikov: I can have a look today, micro-versions and unit tests are the two items on my list
17:17:58 <ildikov> yeah, let me quickly action those
17:18:23 <ildikov> #action ildikov to work on unit tests for the Cinder side changes
17:18:48 <ildikov> #action scottda to put together microversioning for the Cinder side changes
17:18:53 <scottda> cool
17:19:33 <ildikov> #action jgriffith scottda to upload a temporary fix for having the client code used by Devstack on the gate
17:20:06 <ildikov> ok, I think we have the bigger items covered and we can look into a few more after meeting/during the week
17:20:28 <ildikov> jgriffith: anything more for the refactor?
17:23:36 <ildikov> ok, we can move to the testing side
17:23:44 <ildikov> scottda: any updates?
17:24:00 <scottda> I got a review from mriedem on the tempest tests...
17:24:04 <scottda> thanks for that.
17:24:12 <scottda> I've replied and posted a new PS
17:24:31 <scottda> othewise, I've something broken in my job naming...I'll have a look at that next
17:25:08 <scottda> This is a dependency:
17:25:09 <scottda> https://review.openstack.org/#/c/325895/
17:25:21 <scottda> That could use a review, if anyone is up for it.
17:25:55 <mriedem> scottda: mtreinish's comments weren't addressed
17:26:07 <mriedem> i.e. why are we parsing this rather than just passing it through to tempest.conf
17:26:23 <mriedem> you could probably also point out in the review that you're dependent on it
17:26:35 <scottda> mriedem: OK. I hoped the author of the patch would answer that.
17:26:39 <mriedem> and have a change up to project-config for a multi-backend job
17:26:47 <scottda> I thought I did say that in the review...
17:26:50 <mriedem> scottda: if you depend on it, i wouldn't count on others :)
17:27:52 <scottda> OK, well, I'll put it in my queue to deal with that.
17:28:55 <scottda> mriedem: I've this in project-config: https://review.openstack.org/#/c/330678/
17:29:04 <ildikov> scottda: you metioned it, maybe it wasn't direct enough :)
17:29:48 <scottda> ildikov: I guess so.
17:30:24 <scottda> So, I'll circle around on all those patches and get some changes up to address comments, etc.
17:30:37 <mriedem> scottda: more -1 on the project-config change :)
17:31:49 <mriedem> easy stuff
17:32:07 <ildikov> mriedem: I cannot blink as fast as you can comment on and -1 patches :)
17:33:41 <scottda> haha mriedem is fast indeed
17:33:44 <ildikov> scottda: mriedem: IIRC you had a quite long chat around these Tempest and related patches last time. Is there anything remaining we would need to sort out?
17:33:59 <mriedem> i haven't gone back to review the tempest change yet
17:34:23 <scottda> yeah, I need feedback from mriedem first....
17:36:08 <ildikov> ok, then let's continue that chat in the review
17:37:01 <ildikov> #action mriedem to check on scottda's tempest change, #link https://review.openstack.org/#/c/326681
17:37:25 <ildikov> scottda: anything else to this topic?
17:37:54 <scottda> no, I'm good
17:38:06 <ildikov> scottda: cool, thanks for the updates
17:38:47 <ildikov> as the multi-attach work pretty much depends on the Cinder refactor I made only small changes to code for now
17:39:40 <ildikov> like removing some checks in the block_device code etc.
17:40:09 <ildikov> we will only enable multi-attach in Nova for the new code path that uses the new API calls in Cinder
17:40:41 <ildikov> I will also look into the version check code that will be needed on the Nova side as it can be implemented now and then updated later
17:41:56 <ildikov> mriedem: if/when we can get the gate fixed and see the whole flow hopefully working, will there be any reasonable chance to get some review attention this in Nova?
17:42:45 <mriedem> i haven't gone back over john's changes in nova since last week
17:43:05 <mriedem> we're crunching for non-priority FF this thursday so i think people are pushing on things that are closer to making that date
17:43:16 <mriedem> i know i have been
17:43:53 <ildikov> mriedem: the patch is pretty small now, if the base direction looks good to you, then we can work on the micro-version code and ask you to check when the Nova changes is adapted to that
17:44:11 <mriedem> ok
17:44:57 <ildikov> mriedem: I know, but I would not want to actually block the refactor work for that long if there's a chance
17:45:46 <mriedem> so is the question if we're going to land the new cinder api + client + nova changes to use it by thursday?
17:46:42 <mriedem> because i don't think ^ is going to happen
17:46:48 <ildikov> it does not seem realistic until this Thursday
17:47:20 <mriedem> honestly i think a win for newton is getting the new cinder APIs available and tested
17:47:32 <mriedem> and then building on using those in nova in ocata
17:47:45 <mriedem> with the nova dev happening in newton
17:47:53 <mriedem> so we can flush out bugs
17:48:21 <mriedem> cinder APis in newton i mean, assuming those can get in by newton FF on 9/2
17:48:27 <mriedem> i don't know what cinder does about schedules
17:48:51 <ildikov> we still needs some eyes on the Nova side changes as I'm doing my best, but I'm not an expert yet
17:49:02 <mriedem> yeah i realize
17:49:17 <mriedem> i will try to this week, just lots of other people hammering me for things to get in by thursday
17:49:19 <ildikov> thanks for confirming :)
17:50:13 <ildikov> can you check on things after this week as well?
17:51:39 <mriedem> yeah
17:51:49 <ildikov> I just would like to clarify how we can continue after the freeze
17:52:00 <mriedem> it would be good to have some decent idea of the path forward by the nova midcycle
17:52:11 <ildikov> I don't want to argue or anything
17:52:27 <ildikov> yeah, I just wanted ask when you have it
17:52:29 <ildikov> ?
17:52:30 <smcginnis> mriedem: Oh yeah, we should probably try to schedule a cinder-nova midcycle link again.
17:52:32 <mriedem> at the midcycle i'd basically like to say, this is the plan, these are the steps to get us there, and this is what's going ot happen in newton
17:52:42 <mriedem> 7/19-7/21
17:52:47 <ildikov> mriedem: +1
17:53:07 <ildikov> smcginnis: mriedem: we have the two mid-cycles at the same time again, right?
17:53:15 <scottda> ildikov: Yes
17:53:24 <ildikov> scottda: tnx :)
17:53:34 <mriedem> yeah cinder is in ft collins
17:53:39 <mriedem> nova is in hillsboro
17:54:00 <jgriffith> mriedem: ildikov I'm fine with focusing on merging cinder side FWIW
17:54:41 <ildikov> jgriffith: yeap, it would be great if that can happen in Newton
17:55:03 <mriedem> jgriffith: cinder just has a longer lead time since you have until 9/2 i think
17:55:06 <smcginnis> jgriffith: +1
17:55:21 <ildikov> jgriffith: I would like to still start to sort things out on the Nova side as well, so we don't waste too much time on discussions in Ocata
17:55:46 <mriedem> right i think we definitely want to see the nova parts working in the POC in newton
17:55:51 <jgriffith> ildikov: don't get me wrong, I'll still submit, even if it sits and just gets rebased for a few months :)
17:55:52 <mriedem> and tested
17:55:58 <ildikov> smcginnis: I like the Hangouts idea for the mid-cycle as well
17:56:03 <jgriffith> mriedem: +1
17:56:27 <ildikov> jgriffith: I'm experienced in that, so you can count me in ;)
17:56:38 <jgriffith> ildikov: :)
17:56:38 <ildikov> mriedem: +1
17:57:41 <ildikov> #action Cinder team + ildikov to figure out the plan with steps and tasks for Newton until the mid-cycle
17:58:18 <ildikov> #action mriedem + johnthetubaguy to help out on the Nova side if needed
17:58:27 <scottda> BTW, I'm on vacation next week.
17:58:45 <ildikov> #action ildikov + smcginnis to set up a Hangouts call for the mid-cycle
17:58:55 <ildikov> scottda: ok, thanks for the info!
17:59:22 <ildikov> scottda: I'll ping you on Friday for a few updates, if that works
17:59:48 <scottda> ildikov: Sure. Not too late in the MountainDaylight TZ, please :)
18:00:09 <ildikov> scottda: deal :)
18:00:22 <ildikov> is there anything else for today?
18:01:08 <ildikov> ok, if no, then thanks everyone!
18:01:16 <scottda> bye
18:01:32 <ildikov> have a nice day/afternoon :)
18:01:40 <ildikov> #endmeeting