14:01:45 <russellb> #startmeeting nova
14:01:46 <openstack> Meeting started Thu Feb 20 14:01:45 2014 UTC and is due to finish in 60 minutes.  The chair is russellb. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:49 <openstack> The meeting name has been set to 'nova'
14:01:52 <russellb> hello!  who's around?
14:01:58 <mriedem> hi
14:02:03 <BobBall> Mostly.
14:02:14 <llu-laptop> o/
14:02:38 <russellb> 2.5 people!
14:02:39 <russellb> :)
14:02:48 <mriedem> there were at least 3 others in -nova
14:03:01 <russellb> ok, usually a load of lurkers too
14:03:03 <russellb> onward
14:03:06 <russellb> #topic general
14:11:07 <mriedem> ok
14:11:07 <mriedem> in general, i'm all ears
14:11:07 <mriedem> did we lose russellb?
14:11:23 <russellb> #link https://wiki.openstack.org/wiki/Icehouse_Release_Schedule
14:11:23 <russellb> icehouse-3 closing in fast
14:11:23 <russellb> we've past the deadline for having blueprint code proposed
14:11:23 <russellb> that was tuesday of this week
14:11:23 <russellb> feature freeze is march 4
14:11:23 <russellb> < 2 weeks away
14:11:23 <russellb> so we should focus on merging blueprints
14:11:23 <russellb> any questions on the schedule?
14:11:23 <russellb> #topic sub-teams
14:11:24 <russellb> the bot is ignoring me, or we're having some irc troubles ...
14:11:24 <russellb> is this thing on?
14:11:24 * russellb taps on the mic
14:11:24 <annegentle> ow my ears
14:11:24 <annegentle> russellb: I'm seeing topics set and bot stuff
14:11:24 <russellb> seems i'm back now
14:11:32 <mriedem> wow
14:11:33 <russellb> what's the last thing you saw from me ...
14:15:30 <zhangleiqiang> ~
14:15:30 <annegentle> russellb: you tapped the mic
14:15:30 <russellb> heh before that
14:15:30 <mriedem> russellb: the general topic tag
14:15:30 <mriedem> but everything just dumped in
14:15:30 <mriedem> reading scrollback
14:15:49 <garyk> hi
14:15:52 * johnthetubaguy got IRC connected again, and crashes through the door a bit late
14:15:53 <garyk> is this the nova meeting?
14:15:56 <llu-laptop> yes, already started with the topic #general
14:15:56 <llu-laptop> though waiting for your guys to fill the silence here
14:15:59 * johnthetubaguy wonders if IRC is acting funny for other people too
14:16:00 <BobBall> There was a netsplit a short while ago
14:16:00 <BobBall> so probably
14:16:00 <garyk> i can hear silence
14:16:00 <BobBall> In fact only 10 minutes ago.
14:16:00 * llu-laptop sometimes see IRC messages got jammed
14:16:02 <johnthetubaguy> BobBall: ah yes, I did wonder
14:16:04 * johnthetubaguy raises hand for XenAPI news, but BobBall has all the juicy details
14:16:15 <BobBall> Indeed - but not much point going through it without half the people and possibly without the log bot :)
14:16:15 <johnthetubaguy> agreed
14:16:16 <johnthetubaguy> ah, hello
14:16:16 <BobBall> 14:11 < russellb> what's the last thing you saw from me ...
14:16:16 <johnthetubaguy> It just dumped into my screen, with taps on the mic
14:16:16 <johnthetubaguy> and the icehouse-3 closing in fast stuff
14:16:17 * johnthetubaguy raises hand for XenAPI
14:16:17 <BobBall> Looks like IRC isn't back working yet
14:16:17 <johnthetubaguy> yeah, I guess not
14:16:18 <russellb> ah
14:16:18 <annegentle> topic setting does seem slowwww
14:16:18 <russellb> yes this is it, we're starting a bit slow, irc troubles
14:16:18 <russellb> ok well in generall all i had was the release schedule
14:16:18 <russellb> johnthetubaguy: yes
14:16:18 <russellb> big net split i think
14:16:18 <russellb> #topic https://wiki.openstack.org/wiki/Icehouse_Release_Schedule
14:16:18 <russellb> #undo
14:16:18 <russellb> #link https://wiki.openstack.org/wiki/Icehouse_Release_Schedule
14:16:18 <russellb> so, just wanted to make sure everyone is aware that feature freeze is just under 2 weeks away
14:16:18 <russellb> and on to sub-teams ...
14:16:18 <russellb> johnthetubaguy and BobBall: go for it
14:16:18 <mriedem> russellb: you're onto sub-teams now
14:16:18 <mriedem> johnthetubaguy: let's wait for ack from russellb, i think he's gone again
14:16:18 <BobBall> Based on russellb's comments and mriedem's later comments than ours, this is a 3-way split! the most fun kind.
14:16:18 <BobBall> https://wiki.openstack.org/wiki/XenServer/XenServer_CI
14:16:18 <johnthetubaguy> wow
14:16:18 <BobBall> CI is there, processing changes, and was voting for a short time...
14:16:19 <russellb> omg, now i just got a huge flood
14:16:20 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x2e07d50>
14:16:35 <BobBall> I've disabled voting now because two things changed - one was a d-g change in how tempest concurrancy was calculated
14:16:37 <mriedem> are we back?
14:16:38 <BobBall> but we've worked arund that
14:16:46 <BobBall> the second was something we're still digging in to
14:17:00 <BobBall> the result of the second change is that now all tempest full's are failing with different reasons
14:17:00 <johnthetubaguy> mriedem: not sure we are, but this bit is quite one way, so lets go for it
14:17:10 <mriedem> sure
14:17:33 <BobBall> So far all of the failures we've investigated have been traced back to existing bugs - but it's useless to report all of the failures yet
14:17:35 <sdague> internet sneaker chat....
14:17:37 <johnthetubaguy> BobBall: whats the symptom of the second change?
14:17:46 <mriedem> sdague lurks...
14:17:57 <garyk> anyone out there?
14:18:10 <mriedem> garyk: yes, xenapi CI status is going on
14:18:25 <BobBall> The symptom is just that full tempest used to pass now it consistently fails with different bugs which all seem to be races.  We think it's highly likely to be a configuration thing - but haven't tracked it down.
14:18:41 <johnthetubaguy> BobBall: gotcha, ouch
14:18:46 <mriedem> consistently racey
14:18:53 <garyk> BobBall: nice to hear that the CI is up and running
14:18:54 <BobBall> in different races mriedem :D
14:18:56 <johnthetubaguy> yeah, it wasn't so bad yesterday mind
14:19:03 <johnthetubaguy> funky
14:19:15 <garyk> minesweeper also has discovered all kinds of races - bugs atcually..
14:19:19 <BobBall> We don't seem to be getting lucky and getting through them all :/
14:19:19 <mriedem> BobBall: johnthetubaguy: could the known buggy tests be excluded for a time to at least stabilize your CI?
14:19:22 <sdague> BobBall: so based on the fails that you were seeing, I think if you disable volumes tests, all you'd need to do is sort out the issue with that one image test
14:19:38 <mriedem> what sdague said
14:19:48 <johnthetubaguy> yeah, seems like a good path
14:19:53 <BobBall> That's part of the fun... we're using d-g with tox which doesn't have a way to exclude tests ATM
14:20:07 <mriedem> the docker guys were having the same kinds of problems at the meetup last week, high failure rates b/c they were just flooded with fails - they should pair it down and get to a working baseline, then expand
14:20:08 <sdague> BobBall: you can disable volumes tests with a tempest config variable
14:20:15 <BobBall> oh brilliant sdague !
14:20:21 <BobBall> We'll do that then.
14:20:30 <BobBall> and see what falls out
14:20:37 <mriedem> yeah, start small
14:20:46 <mriedem> otherwise it's overwhelming
14:20:48 <BobBall> matel is also looking at adding a specific exclude list
14:21:04 <BobBall> basically discovering all tests then removing some from the list before getting testr to run them
14:21:13 <BobBall> but disabling volume tests is easy to do
14:21:29 <mriedem> BobBall: do you guys use nova-network?
14:21:32 <BobBall> So I'll do that and we'll see where we get to later today when it's ran a fwe tests.
14:21:35 <BobBall> yes, nova-network
14:22:18 <johnthetubaguy> BobBall: you happy with that plan? any more updates?
14:22:33 <BobBall> No more updates
14:22:34 <russellb> very happy to see the progress
14:22:35 <johnthetubaguy> BobBall: do you have the capacity you need to drain the queue now?
14:22:55 <BobBall> Probably john
14:23:22 <mriedem> other sub-team status?
14:23:23 <johnthetubaguy> BobBall: cool, more than you had yesterday at least, thats the main
14:23:30 <garyk> i have a concern regarding all of the third party CI's - when there is the end of cycle rush - it is going to be intersting to see how it all works.
14:23:38 * johnthetubaguy raises scheduler hand if none else
14:23:50 <russellb> johnthetubaguy: sure, update on your fun scheduler?
14:23:56 <russellb> scheduler optimization at least
14:24:04 <johnthetubaguy> oh right, I wasn't going to talk too much about that
14:24:08 <russellb> oh :)
14:24:12 <russellb> thought that's why you raised your hand
14:24:15 <johnthetubaguy> but anyways, a little bit of DB caching is going a long way
14:24:18 <johnthetubaguy> please review that
14:24:21 <russellb> heh
14:24:28 <russellb> yeah, seems doable for icehouse
14:24:29 <johnthetubaguy> no, was at the meeting yesterday
14:24:33 <johnthetubaguy> talk about a split
14:24:34 <russellb> cool
14:24:37 <johnthetubaguy> well, I suggested one at least
14:24:42 <russellb> meeting split?
14:24:46 <russellb> or?
14:24:48 <johnthetubaguy> just finding the dam etherpad
14:24:54 <johnthetubaguy> oh, scheduler split, not meeting split
14:24:58 <russellb> ah
14:25:20 <johnthetubaguy> #link https://etherpad.openstack.org/p/icehouse-external-scheduler
14:25:28 <russellb> yeah, and consensus seems to be that even if we don't think gantt is close to ready, we should leave it in openstack/ and not move to stackforge/
14:25:33 <russellb> because moving repos around is a PITA
14:25:48 <johnthetubaguy> ah thats cool
14:25:57 <johnthetubaguy> we would only have to move it again
14:26:01 <johnthetubaguy> so, the plan
14:26:08 <johnthetubaguy> its at the bottom of that etherpad
14:26:13 <johnthetubaguy> basically, like no-db-compute
14:26:15 <garyk> russellb: will there be time set aside at the summit to speak about gantt?
14:26:17 <russellb> plan b?
14:26:25 <garyk> it seems like a topic on its own
14:26:31 <russellb> garyk: as a part of the nova schedule
14:26:42 <johnthetubaguy> lets separate out obvious bits, then see where we are, basically
14:26:42 <russellb> a project has to be incubated to get its own time
14:26:58 <garyk> i think at least one nova session should be devoted to it and then it may split out onto its own track
14:27:14 <russellb> yeah when we originally talked about this, i think i was assuming we'd have the conductor bits done
14:27:17 <garyk> ok, it will be an incubated project.
14:27:46 <johnthetubaguy> russellb: yeah, sorry about that, it sucks
14:27:53 <russellb> oh no worries
14:27:58 <russellb> just a bit of dependency management fail
14:28:17 <johnthetubaguy> yeah, it should have been higher priority sooner
14:28:19 <johnthetubaguy> oops
14:28:37 <russellb> but all good, on from here
14:28:49 <russellb> but the plan does sound reasonable
14:28:52 <johnthetubaguy> if we get time, should we push that through?
14:28:58 <johnthetubaguy> the conductor bits that is
14:29:02 <russellb> sure, would love to if we can
14:29:12 <johnthetubaguy> yeah, thinking RPC cleanup
14:29:15 <russellb> exactly
14:29:27 <russellb> if we do it before release, we can remove all that code once juno opens
14:29:32 <russellb> otherwise it's stuck until K
14:29:32 <johnthetubaguy> right, well its bumped a bit higher in a massive pile of doom
14:30:04 * johnthetubaguy feels he is done typing for now, wonders if we will cover v3 API
14:30:24 <russellb> yes, we need to cover that
14:30:41 <russellb> maybe in blueprints
14:30:46 <russellb> want to talk bugs real quick, then to blueprints
14:30:48 <russellb> #topic bugs
14:31:07 <russellb> so, tracy jones has taken the bug czar role now
14:31:14 <johnthetubaguy> yeah, bug sub groups seems to be forming now :)
14:31:18 <russellb> she put out a message to openstack-dev trying to get an initial team together
14:31:29 <garyk> saldy there are not too many volutneers
14:31:34 <russellb> #link http://lists.openstack.org/pipermail/openstack-dev/2014-February/027190.html
14:31:46 <russellb> more now
14:31:51 <russellb> but more the merrier :-)
14:31:54 <garyk> nice to hear
14:32:05 <russellb> really need a solid group that can devote an hour to this every week
14:32:09 <russellb> at least an hour is all i ask :-)
14:32:14 <johnthetubaguy> List is not too tiny right now https://etherpad.openstack.org/p/nova-bug-management
14:32:34 <johnthetubaguy> that seems like a good plan, an hour a week
14:32:38 <russellb> so, if you're interested, please sign up!
14:32:51 <russellb> i'm quite hopeful that this renewed effort will help improve our bug queue
14:33:06 * mriedem wishes the ML thread was tagged with [nova], never saw it
14:33:20 <russellb> after feature freeze, we'll all switch focus to bugs, so more on bugs in a couple weeks ...
14:33:25 <russellb> #topic blueprints
14:33:35 <russellb> and speaking of czars, we have a blueprint czar!
14:33:40 <johnthetubaguy> #link https://etherpad.openstack.org/p/nova-icehouse-blueprint-cull
14:33:41 <russellb> johnthetubaguy: <--- \o/
14:33:44 <mriedem> russellb: i did have a bugs topic in the agenda, we can do in open discussion
14:33:50 <russellb> mriedem: ah sorry
14:34:04 <mriedem> let's come back to it
14:34:12 <mriedem> later i mean :)
14:34:23 <garyk> i am not sure how anmial rights groups are going to deal with the cull
14:34:28 <johnthetubaguy> so lots of shifting blueprints into Juno, those without all their patches up
14:34:46 <russellb> and i probably didn't do a great job of recording all the changes i made on the etherpad, sorry
14:34:53 <garyk> the deadline for code in review was the 18th. is that correct?
14:35:02 <russellb> garyk: correct
14:35:12 <russellb> so, v3 blueprints ...
14:35:14 <johnthetubaguy> russellb: yeah, hey ho, me neither
14:35:46 <russellb> #link http://lists.openstack.org/pipermail/openstack-dev/2014-February/027588.html
14:36:00 <johnthetubaguy> yeah, we have two
14:36:08 <johnthetubaguy> I mean, to
14:36:20 <russellb> there are a bunch of v3 blueprints, what should we do with them?
14:36:25 <russellb> we previously had them prioritized up to Medium
14:36:31 <russellb> but given the post above, i dropped them to Low
14:36:31 <johnthetubaguy> well, low for sure
14:36:36 <mriedem> +1
14:36:46 <johnthetubaguy> I wonder about defer to juno though
14:36:48 <russellb> should we leave them and just let whatever happens to merge get in?
14:36:56 <johnthetubaguy> yeah, duno, to be honest
14:37:05 <russellb> or actively defer things, even like this, to help focus on a smaller set?
14:37:07 <johnthetubaguy> we could do with a bit more review bandwidth for the other stuff
14:37:14 <russellb> yeah, that's what i'm worried about
14:37:20 <sdague> well, I was wondering about the meta issue. Because it seems like we should figure out if we're still going ahead on the v3 front as planned
14:37:22 <johnthetubaguy> yeah, its tempting, feels bad not asking Chris though
14:37:37 <russellb> yeah, should definitely talk to him
14:37:42 <russellb> bummer he's not here
14:37:44 <johnthetubaguy> sdague: I think its a question of how to maintain v2
14:37:50 <johnthetubaguy> yeah, stupid timezones
14:37:52 <sdague> and if not, I really wonder about any review time on it. That also includes the api validation bits
14:38:03 <russellb> i haven't had a chance to read your post sdague , sorry
14:38:07 <sdague> because there are a bunch of other things that are only happening on v3
14:38:28 <mriedem> johnthetubaguy: 'how to maintain v2' is the same topic as my bug item...
14:38:32 <johnthetubaguy> oh IRC just broke again :(
14:38:38 <johnthetubaguy> mriedem: its a big fella
14:38:40 <sdague> yeh, silly irc
14:38:55 <russellb> really messing up our meeting today
14:39:07 <russellb> ok, so let's see if there's anything else for today, and if not, we can use the rest of the time on the API topic
14:39:16 <russellb> johnthetubaguy: any other thoughts on blueprints?
14:39:23 <johnthetubaguy> yeah, not much really...
14:39:29 <johnthetubaguy> please keep them up to date, I guess
14:39:41 <johnthetubaguy> but I will try keep people honest on that, once we get the list down
14:39:43 <mriedem> i had a general discussion item
14:39:46 <russellb> and we should try to put more review focus on the targeted blueprints that have all code up
14:39:55 <johnthetubaguy> I might start un-approving things too, that are stale
14:39:56 <johnthetubaguy> etc
14:40:04 <russellb> johnthetubaguy: +1
14:40:06 <mriedem> johnthetubaguy: that's a good idea
14:40:15 <russellb> johnthetubaguy: i found some that had code, but have been abandoned for 1 or 2 months ... just deferred
14:40:20 <johnthetubaguy> yeah, I will try get a list together for the patches too, just one tiny error
14:40:28 <johnthetubaguy> russellb: likewise, I will try clean that up
14:40:41 <russellb> johnthetubaguy: really appreciate the help, it's no small pile of work
14:40:50 <russellb> cheers for our blueprint czar!  :-)
14:40:53 <johnthetubaguy> I am seeing how much time it takes, no worries
14:40:57 <mriedem> furry hats for everyone!
14:40:58 <johnthetubaguy> anyways, that all for now
14:41:01 <johnthetubaguy> hehe
14:41:14 <russellb> #topic open discussion
14:41:24 <russellb> mriedem: what's up
14:41:27 <johnthetubaguy> mriedem: http://www.youtube.com/watch?v=fWucPckXbIw&feature=kp
14:41:36 <mriedem> i was just going to bring up jog0's post last night about an oslo sync tema
14:41:38 <mriedem> *team
14:41:42 <mriedem> he's not here i don't think though
14:41:50 <mriedem> was wondering what the feelings are on that
14:41:56 <mriedem> since oslo sync seems to be a contentious issue
14:41:58 <russellb> #link http://lists.openstack.org/pipermail/openstack-dev/2014-February/027631.html
14:42:22 <johnthetubaguy> I don't like the risk right now, we should do it when Juno opens for sure
14:42:32 <johnthetubaguy> but maybe we should take the hit anyway, it could fix things
14:42:39 <johnthetubaguy> its a nice fence I am on here
14:42:46 <mriedem> johnthetubaguy: yeah, i think jog0 would say we want to sync before icehouse closes to get any bug fixes
14:42:48 * russellb joins johnthetubaguy on the fence
14:42:49 <mriedem> but yeah, i see both ways
14:42:56 <mriedem> make room
14:43:11 <sdague> honestly, if we think we have enough time to catch fallout, I'd say damn the torpedos on this one
14:43:29 <mriedem> at a high level, it seems to be "well if we exclude this and this and this module, auto-sync should be fine"
14:43:29 <sdague> because we've definitely run into bugs in the past where were entirely because olso was old
14:43:40 <ndipanov> sync, this is the best time to do it I'd say
14:43:45 <russellb> so basically the assertion is that ad-hoc syncs aren't good enough, and someone should just take ownership of it?
14:43:51 <russellb> i at least agree that sure, we should sync
14:43:53 <johnthetubaguy> lets just sync the lot, and if the gate breaks, revert it?
14:44:03 <russellb> +1 :)
14:44:09 <sdague> :)
14:44:10 <johnthetubaguy> russellb: by Juno-1 we should be in sync
14:44:18 <ndipanov> yeah revers are easy no one touches that code except syncs
14:44:26 <ndipanov> +1
14:44:31 <mriedem> well, and jog0 is doing a full sync on cinder as a POC
14:44:39 <mriedem> so he'll probably learn what to trim out there first
14:44:56 <russellb> oslo folks also seem to be working hard on splitting that code out into libs now
14:45:08 <mriedem> yeah, that's nice, still some bumps to iron out there
14:45:10 <russellb> they have a big dependency tree between the stuff in there and are working out from the leaves
14:45:14 <mriedem> yup
14:45:16 <mriedem> log_handler
14:45:24 <johnthetubaguy> hmm, there is a thought...
14:45:46 <johnthetubaguy> we should sync before icehouse ends, else its gona make upgrade hard
14:45:52 <russellb> but there's still the icehouse question
14:45:56 <russellb> i'm all for syncing for icehouse
14:46:08 <russellb> and regarding his idea, having someone take ownership of making sure we stay up to date, sure
14:46:09 <sdague> +1
14:46:09 <russellb> seems fine
14:46:36 <johnthetubaguy> seriously, lets just see if it breaks the gate or not, then see were we are
14:46:54 <mriedem> i'm sure jog0 will be happy to hear this
14:46:57 <russellb> heh, k :)
14:47:06 <russellb> so we can catch up more later in -nova when he's up
14:47:12 <russellb> so, on to the API stuff?
14:47:15 <mriedem> sure
14:47:17 <johnthetubaguy> yeah
14:47:24 <russellb> "now what"
14:47:31 <mriedem> so my topic...
14:47:40 <mriedem> we have two bugs for v2 regarding neutron,
14:47:50 <mriedem> one for the quotas extension and one for the limits extension,
14:48:07 <mriedem> both have patches, in one we proxy to neutron (limits), in the other (proposed last night) we don't
14:48:13 <mriedem> but both are needed to make the tests work in tempest
14:48:28 <mriedem> so if we're going to live with v2 for awhile, and support neutron, do we proxy or not?
14:48:40 <russellb> wellllll
14:48:44 <mriedem> v3 doesn't proxy, which is fine, but this hodge podge in v2 kind of sucks
14:48:50 <russellb> my general feeling is proxying sucks
14:48:50 <mriedem> at least regarding anything to do with security groups and floating IPs
14:48:55 <johnthetubaguy> depends if info_cache gets updated I guess?
14:49:06 <russellb> for end user facing APIs, i think we have to, for compatibility
14:49:13 <russellb> for admin APIs, i'm less concerned
14:49:15 <sdague> right, but v2 is all about the proxy. We have proxy API for images, volumes, and network
14:49:15 <mriedem> russellb: that was my feeling
14:49:23 <russellb> sdague: right
14:49:29 <mriedem> also note that the limits api change is ready
14:49:30 <russellb> so, i wonder if we should draw the line between cloud user or admin
14:49:32 <mriedem> i've already tested it with tempest
14:49:34 <russellb> quotas?  meh,
14:49:57 <johnthetubaguy> but users list their own quotas, or does it not do that?
14:50:06 <russellb> oh, you're right
14:50:16 <mriedem> in this case i see limits and quotas as being tied together
14:50:29 <mriedem> so if you are going to proxy for one i'd assume (as a user) that you will for the other also
14:50:35 <russellb> limits and quotas sure sound the same don't they.
14:50:58 <mriedem> well, and you find out here how they are tied together: https://review.openstack.org/#/c/74839/
14:51:09 <mriedem> and the ugly hack if they aren't
14:51:10 <johnthetubaguy> so, whats the question again?
14:51:18 <johnthetubaguy> should we proxy for neutron?
14:51:24 <mriedem> the question is for these two bugs for the APIs, do we proxy to neutron or not, yeah
14:51:39 <johnthetubaguy> its consistent at least
14:51:40 <mriedem> also note that in havana, the v2 limits API bug was marked high
14:51:41 <russellb> i suppose the default answer is always "proxy, or whatever we have to do, to keep APIs working, when reasonable to do so"
14:52:03 <russellb> so proxy, unless we have a good excuse not to support it
14:52:08 <johnthetubaguy> russellb: +1 but nothing in v3, whatever that is
14:52:13 <mriedem> ok, agreed
14:52:20 <russellb> yeah, all of that for v2 specifically
14:52:49 <mriedem> i wanted to bring this up specifically for the quotas patch since i -1'ed it for inconsistency
14:52:59 <russellb> thanks :)
14:53:04 <mriedem> that's it for me
14:53:22 <russellb> so sdague, i guess you were bringing up http://lists.openstack.org/pipermail/openstack-dev/2014-February/027688.html
14:53:26 <johnthetubaguy> cool, so general v3 stuff?
14:53:30 <sdague> russellb: yeh
14:54:11 <sdague> because with the idea that v2 is around for a while, I think we need to reassess the gameplan for the nova API
14:54:22 <russellb> agree.
14:54:34 <sdague> with a goal of 1 API with 1 data format in L (or worst case M)
14:54:40 <sdague> and how do we get there
14:54:43 <russellb> so you've jumped right to some really hard questions i was a bit afraid to ask (yet at least), heh
14:54:57 <sdague> well just call me an instigator :)
14:55:13 <johnthetubaguy> I kinda thing we don't like the current format
14:55:25 <johnthetubaguy> so I like the idea to move to a slightly updated one
14:55:27 <russellb> yeah, it hasn't gone how anyone really expected
14:55:35 <johnthetubaguy> if only to add in the tasks support
14:55:49 <russellb> so tasks ... could we do tasks in v2?
14:56:01 <johnthetubaguy> yeah, through headers and stuff, but it sucks
14:56:10 <russellb> does it suck more than 2 APIs?
14:56:20 <johnthetubaguy> fair point
14:56:24 <johnthetubaguy> probably not
14:56:35 <johnthetubaguy> the inconsistencies in v3 would be nice to remove
14:56:37 <russellb> not trying to imply an answer, just asking, it's what we have to figure out
14:56:45 <johnthetubaguy> if only by some filter that accepts both
14:56:53 <sdague> we also said we don't like proxy APIs, but if we are talking about v2 largely forever, we need a mechanism to deprecate and remove those
14:57:02 <johnthetubaguy> that would really reduce the testing though?
14:57:14 <johnthetubaguy> we could unit test that stuff, and leave tempest on the new one
14:57:25 <johnthetubaguy> the return values are harder right
14:57:28 <russellb> the more I think about what major API revs mean, the more I think it should be a fundamental do-over almost
14:57:29 <sdague> because the internals on v2 are way more gross and fragile than 3
14:57:33 <johnthetubaguy> and deprecating XML is harder
14:57:56 <russellb> on internals, can we backport internals improvements to v2?
14:58:11 <sdague> russellb: it's not so much a backport as a gut
14:58:25 <russellb> re-do the internals improvements on v2  :-/
14:58:35 <johnthetubaguy> possibly, but versioning extensions, we could add that at some point I guess
14:58:56 <russellb> not sure why we couldn't
14:59:03 <sdague> it's all just time and motivation
14:59:09 <johnthetubaguy> its only the XML clients, and well...
14:59:28 <sdague> and I think a big piece is first agreeing where we are headed
14:59:34 <johnthetubaguy> I guess we are out of time
14:59:42 <russellb> let's continue on the ML
14:59:44 <johnthetubaguy> sdague: +1
14:59:46 <sdague> because 1 API with 1 data format I think is an important thing to see if we agree on
15:00:00 <sdague> agreed, we can take to the ML
15:00:06 <russellb> thanks everyone
15:00:13 <russellb> #endmeeting