17:00:02 <ildikov> #startmeeting cinder-nova-api-changes
17:00:03 <openstack> Meeting started Mon Aug  8 17:00:02 2016 UTC and is due to finish in 60 minutes.  The chair is ildikov. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:06 <openstack> The meeting name has been set to 'cinder_nova_api_changes'
17:00:10 <ildikov> scottda ildikov DuncanT ameade cFouts johnthetubaguy jaypipes takashin alaski e0ne jgriffith tbarron andrearosa hemna erlon mriedem gouthamr ebalduf patrickeast smcginnis diablo_rojo gsilvis
17:00:20 <jgriffith> ildikov: hi
17:00:31 <ildikov> jgriffith: hi :)
17:01:22 <jgriffith> ildikov: I've got the shelved instance attach/detach issues fixed up
17:01:43 <hemna> hey
17:01:51 <jgriffith> ildikov: I have introduced some other things that need cleaning but I should have a working tempest run version of the code shortly for each project
17:01:59 <hemna> jgriffith, https://review.openstack.org/#/c/257853/
17:02:06 <ildikov> jgriffith: that's great!!!\
17:02:14 <jgriffith> ildikov: I had hoped to push that update over the week-end, but alas didn't get to it... travelling the next few days
17:02:16 <hemna> worked with matt to get that to land last week
17:02:37 <jgriffith> hemna: yeah, saw that
17:02:41 <jgriffith> hemna: cool
17:02:48 <jgriffith> hemna: I guess I might need to change what I did
17:03:05 <hemna> nova was calling terminate_connection with a completely bogus/faked connector
17:03:06 <jgriffith> hemna: I just push an attachment_id into the connection_data field
17:03:13 <jgriffith> hemna: yeah, I"m aware :)
17:03:17 <hemna> hehe
17:03:20 <jgriffith> hemna: but it sort of "had to"
17:03:28 <hemna> so, now they shouldn't be calling terminate_connection in this case.
17:03:49 <jgriffith> hemna: in the case of shelved, although it would've been nice if a little effort went in to creating an API semantic for it :)
17:04:09 <hemna> we can always change that
17:04:31 <hemna> this was kinda of an emergency as lots of cinder driver CIs were failing 100% as well as brick CIs
17:04:44 <ildikov> jgriffith: those parts I meant need a fix when the basic attach/detach work with your new API
17:04:52 <hemna> ildikov, +1
17:05:07 <jgriffith> hemna: http://paste.openstack.org/show/552062/
17:05:20 <jgriffith> hemna: not quite as elegant as what you linked to :)
17:05:50 <ildikov> although at least there are some tests failing, so we have some coverage...
17:05:55 <jgriffith> hemna: and I happily *ignore* the current stuff :)
17:06:02 <hemna> hehe
17:06:14 <scottda> oops. Late again...
17:06:56 <jgriffith> that diff doesn't include shoving the attachment_id in during attach, but that's in my github if you wanted to see it (at least I think it is updated) :)
17:07:11 <jgriffith> and if not it will be soon :)
17:07:15 <hemna> scottda, no soup for you!
17:07:24 <ildikov> jgriffith: 'supports_new_attach' is the flag we talked about and then we will use microversion, right?
17:07:40 <scottda> I need to get this meeting into Outlook...
17:07:49 <jgriffith> ildikov: yeah
17:07:58 <ildikov> scottda: :)
17:08:05 <ildikov> jgriffith: ok, cool
17:08:10 <jgriffith> scottda: what's Outlook?
17:08:16 <ildikov> hemna: LOL :)
17:08:28 <scottda> It's this thing my company makes me use....
17:08:51 <ildikov> scottda: I used to know that feeling
17:09:46 <ildikov> and then I changed company :)
17:10:41 <ildikov> jgriffith: I added some hack to use the regular policy file in the attach test: https://review.openstack.org/#/c/339692/6
17:11:06 <hemna> jgriffith, so is there anything I can help with, while I'm still here at HPE.  I'm trying to stay busy until I'm out of a job.
17:13:19 <cFouts> o/
17:14:49 <jgriffith> hemna: for sure
17:14:58 <jgriffith> hemna: let me get it stable and then we can attack
17:15:14 <jgriffith> hemna: the biggest thing is probably going to be micro-versions, I want nothing to do with that :)
17:15:24 * mriedem joins late
17:15:49 <hemna> hehe ok
17:16:15 <ildikov> jgriffith: I think I saw scottda volunteering for that ;)
17:16:48 <scottda> so, I'm wondering at what point we want the microversion support...
17:16:54 <scottda> because it will make testing harder.
17:17:09 <scottda> The Nova side of the micrsoversion support will need to be in place once Cinder has it.
17:17:13 <ildikov> mriedem: jgriffith is working on the fixes for shelve and also on tempest tests
17:17:38 <mriedem> ildikov: ok
17:18:07 <ildikov> scottda: yeap and there will be quite some if/else in Nova
17:18:38 <scottda> Yeah, so we can move more quicly on this POC stuff without microversions...
17:18:51 <ildikov> mriedem: I also more or less fixed the remove check_attach patch: https://review.openstack.org/#/c/335358/
17:19:26 <scottda> When I get around to it, I'll post the microversion support in a separate patch with a dependency. That way it'll be available to work on the nova plumbing, but it won't block dev/testing on the stuff jgriffith is working on.
17:19:28 <ildikov> mriedem: there's a test failure that pops up, but only randomly: http://logs.openstack.org/58/335358/11/check/gate-nova-python34-db/cc3c89f/console.html#_2016-08-07_17_02_54_790835
17:19:47 <ildikov> mriedem: do you happen to have an idea what causes this issue?
17:19:48 <mriedem> ildikov: mox tests are racy under py3
17:19:59 <mriedem> would have to be converted to mock
17:20:23 <ildikov> mriedem: ah, ok
17:20:38 <ildikov> mriedem: is it ok in a follow up patch?
17:20:43 <mriedem> or you have a global state race in the unit testes
17:20:57 <mriedem> ildikov: i'd probably do it as a patch that comes before the change actually
17:21:47 <ildikov> mriedem: ok, I'll do that then, I don't think it's a global state, there's nothing in the reserve volume test as far as I can see, but I might be mistaken
17:21:59 <ildikov> scottda: that sounds like a plan!
17:23:33 <ildikov> mriedem: can you check the patch, whether there's anything missing from the version bumps, etc?
17:23:46 <mriedem> at some point yet
17:23:48 <mriedem> *yeah
17:24:10 <ildikov> ok, tnx :)
17:25:15 <ildikov> jgriffith: BTW are you fully off this week due to travels?
17:26:26 <jgriffith> ildikov: nope
17:26:36 <jgriffith> Just off and on for a few days
17:27:11 <ildikov> ok cool, tnx for the info
17:27:17 <jgriffith> sure
17:28:37 <ildikov> hemna: I started to look into how to fix some issues around the functional tests in jgriffith's cinder patch
17:29:13 <ildikov> hemna: if you have some spare time and you can look into that I can always use some help to figure out testing in Cinder :)
17:32:56 <ildikov> mriedem: also thanks for the Devstack gate patch!
17:34:21 <ildikov> anything else from anyone for today?
17:38:04 <ildikov> ok then let's continue on the reviews, thanks everyone!
17:38:43 <ildikov> #endmeeting