16:00:09 <ildikov> #startmeeting cinder-nova-api-changes
16:00:10 <openstack> Meeting started Thu Aug 31 16:00:09 2017 UTC and is due to finish in 60 minutes.  The chair is ildikov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:13 <openstack> The meeting name has been set to 'cinder_nova_api_changes'
16:00:13 <mriedem> o/
16:00:21 <smcginnis> o/
16:00:30 <ildikov> johnthetubaguy jaypipes e0ne jgriffith hemna mriedem patrickeast smcginnis diablo_rojo xyang1 raj_singh lyarwood jungleboyj stvnoyes
16:00:55 <ildikov> let's wait a minute more and then dive in
16:01:00 <jungleboyj> @!
16:01:01 <_pewp_> jungleboyj |。･ω･|ﾉ
16:01:33 <stvnoyes> o/
16:01:53 <xyang1> Hi
16:02:03 <ildikov> ok, let's start :)
16:02:18 <ildikov> the open reviews are here: https://review.openstack.org/#/q/topic:bp/cinder-new-attach-apis+status:open
16:02:28 <ildikov> smcginnis: I payed attention to 'open' this time :)
16:02:45 <ildikov> jungleboyj: did you get there yesterday to cut a new client release?
16:03:10 <jungleboyj> ildikov:  I did.  I haven't checked to see if it went through.
16:03:39 <jungleboyj> ildikov:  https://review.openstack.org/#/c/499282/
16:03:41 <jungleboyj> Not yet.
16:04:16 <ildikov> ok, I guess it shouldn't take that long
16:04:55 <jungleboyj> smcginnis:  Anything you can do to get that through?
16:05:04 <ildikov> so until we don't have the new client we still won't see clean test runs on the gate, but we're pretty close now
16:06:17 <ildikov> mriedem: I proposed the spec for Queens
16:06:36 <ildikov> mriedem: also commented on the begin_detaching one
16:06:49 <mriedem> ok
16:07:17 <smcginnis> ildikov, jungleboyj: I'll push that through. No need to hold off at this point.
16:07:31 <ildikov> smcginnis: sounds good, thanks!
16:07:40 <ildikov> smcginnis: do we need a g-r patch too?
16:08:26 <smcginnis> ildikov: The new package will automatically generate an upper-constraints update, but yes, we will need to raise the minumum g-r if we want to make sure the right version is always used.
16:08:48 <ildikov> smcginnis: ok, got it
16:09:01 <ildikov> smcginnis: jungleboyj: will either of you handle that?
16:09:36 <smcginnis> I can if jungleboyj doesn't jump on it.
16:09:46 <jungleboyj> smcginnis:  Another learning op for me?
16:10:02 <smcginnis> jungleboyj: Sure, we can work on that together.
16:10:26 <jungleboyj> smcginnis:  Thanks.  All new to me.
16:10:42 <ildikov> jungleboyj: smcginnis: cool, thanks!
16:11:18 <ildikov> beyond this we have a bug report from stvnoyes on when to have the attachment_id appearing in the volume details
16:11:51 <mriedem> jungleboyj: bumping a minimum version in g-r is new to you after all these years?!
16:12:03 <ildikov> as currently we only have attachment_id when the attachment is crated and connected on both sides
16:12:13 <jungleboyj> Oh, in global-requirements.
16:12:16 <jungleboyj> I guess not.
16:12:20 <smcginnis> mriedem: Looks like this should probably just be abandoned: https://review.openstack.org/#/c/466184/
16:12:23 <jungleboyj> Now that I think about it.
16:12:47 <stvnoyes> ildikov- that bug was addressed by this - https://review.openstack.org/#/c/498055/ and was merged
16:12:48 <jungleboyj> Let me go tackle it.  Just been a while.
16:13:00 <mriedem> smcginnis: i think i already -2d or abandoned another similar one for mitaka from him
16:13:05 <mriedem> someone is playing the backport fork game
16:13:09 <ildikov> smcginnis: I asked Ashish in Boston to upload their code to see whether we can reuse anything for multi-attach
16:13:17 <smcginnis> mriedem: Annoyance level high.
16:13:21 <ildikov> smcginnis: mriedem: it's not for merge
16:13:26 <mriedem> i realize
16:13:28 <mriedem> but
16:13:32 <mriedem> it's not helpful either
16:13:47 <smcginnis> jungleboyj: We'll need the release go through first, then we can make the g-r bump Depends-On the upper-constraints patch.
16:13:56 <ildikov> smcginnis: mriedem: my bad, next time I will tell people to use an own repo for this purpose
16:14:26 <ildikov> mriedem: and yeah, I didn't know in advance whether it will be or not... :)
16:14:45 <ildikov> the patch itself can be abandoned
16:15:10 <jungleboyj> smcginnis:  Ok.
16:15:27 <ildikov> mriedem: I thought to look into the old multi-attach spec and re-propose an updated version for Queens
16:15:35 <ildikov> mriedem: if you're not against the idea
16:16:29 <mriedem> ildikov: that's fine
16:16:52 <ildikov> mriedem: ok
16:18:35 <ildikov> stvnoyes: but you reported another bug since that
16:19:23 <ildikov> stvnoyes: with attachment_id not in volume details until attachment_complete is called
16:19:28 <stvnoyes> that was https://bugs.launchpad.net/cinder/+bug/1713521 - where the attachment disappears after update but before complete
16:19:29 <openstack> Launchpad bug 1713521 in Cinder "v3 attach, volume shows no attachments for a short period during migrate" [Undecided,Confirmed]
16:19:52 <stvnoyes> not the actual attach, but how it is reported
16:20:34 <stvnoyes> is anyone looking at that? I thought jgriffith would be interested, but it seems he's not around?
16:20:51 <ildikov> stvnoyes: well, it's not there until complete, which means that for live_migrate there's a window where you don't have that data in the volume details
16:20:59 <stvnoyes> correct
16:21:24 <stvnoyes> i would guess it also happens during attach, not just migrate (but I haven't checked that)
16:21:55 <stvnoyes> if john's away, I'll take a look into it
16:23:00 <ildikov> I think it's a bit of a philosophical question as well on when we want that data to be present
16:24:01 <ildikov> as we're mostly using the volume ref on the Nova side I think it would make sense to have it listed from the moment of reserving the volume
16:24:45 <ildikov> does anyone have a strong opinion on this?
16:25:35 <stvnoyes> ok, let me see why it's doing that. then we can have a better idea if what we should do
16:25:48 <stvnoyes> if=of
16:26:07 <ildikov> stvnoyes: sounds good, thanks!
16:26:37 <ildikov> that was it from me for today
16:26:57 <ildikov> I hope the client gets through and we can get the reviews move forward too
16:27:56 <ildikov> would be great to have this in shape by the PTG and merge things there the latest
16:28:05 <ildikov> so we can move to the multi-attach bits
16:29:48 <ildikov> ok, I think that's it for today then :)
16:30:15 <ildikov> thank you all
16:30:19 <smcginnis> Thanks ildikov!
16:30:28 <ildikov> keep in touch on the channels till next week! :)
16:30:38 <stvnoyes> kk
16:30:46 <ildikov> #endmeeting