15:00:40 <n0ano> #startmeeting gantt
15:00:41 <openstack> Meeting started Tue Mar 24 15:00:40 2015 UTC and is due to finish in 60 minutes.  The chair is n0ano. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:45 <openstack> The meeting name has been set to 'gantt'
15:00:54 <n0ano> anyone here to talk about the scheduler?
15:01:45 <alex_xu> o/
15:01:59 <edleafe> o/
15:02:03 <markus_z> o/
15:03:03 <n0ano> let's get started then...
15:03:25 <bauzas> \o
15:03:30 <n0ano> #topic Patch status - https://etherpad.openstack.org/p/kilo-nova-priorities-tracking
15:03:46 <edleafe> I have nothing to talk about for once
15:03:47 <n0ano> edleafe, you're off the hook, all your patches got merged :-)
15:03:53 <edleafe> :)
15:04:25 <bauzas> I don't have much to say except I'm on my way too :)
15:04:37 <n0ano> I am curious about Paul's, 4 of his got merged yesterday, is there hope for the remaining 5?
15:05:36 <bauzas> n0ano: which ones got merged ?
15:05:53 <n0ano> bauzas, check the priority page, the last 4 on his list
15:06:34 * n0ano is still a little fuzzy on what can/cannot get merged after a freeze
15:06:34 <bauzas> n0ano: oh ok, testing stuff
15:06:44 <bauzas> n0ano: :)
15:06:45 <lxsli> o/
15:06:50 <edleafe> n0ano: is there a FFE for these?
15:07:06 <sean-k-mooney> quit
15:07:25 <lxsli> n0ano: my patches got merged without any FFE (I know of) so a) yay!, b) not sure what's going on there
15:07:26 <n0ano> I don't remember seeing one, I was letting Paul drive that
15:07:49 <lxsli> think he plans to rebase them and cross fingers for +W
15:08:08 <n0ano> Paul's patches are in the priority list, so maybe that means it gets an extension by definition
15:08:16 <edleafe> I haven't seen Paul around much
15:08:18 <lxsli> ah could be it, mine were too
15:08:19 <edleafe> I can rebase them
15:08:29 <lxsli> he has a clash right now, let me ask him
15:08:38 <edleafe> n0ano: I wouldn't assume that
15:08:58 <edleafe> johnthetubaguy was pretty concerned about my patches barely making it *before* FF
15:09:00 <bauzas> n0ano: honestly, don't brainstorm too much about that
15:09:32 <n0ano> bauzas, meaning just let the process work?
15:09:46 <bauzas> n0ano: meaning let the discussions happen on IRC :)
15:10:41 <lxsli> edleafe: ok he's talking atm, can't imagine he'll mind though! Thank you
15:10:42 <n0ano> well, my attitude is keep working things until someone does a -2 and officially moves it to the next release
15:11:13 <lxsli> n0ano: +1, no harm done by keeping up to date
15:11:48 <n0ano> rebasing would be good, edleafe can I have you work out who will do it?
15:11:57 <edleafe> n0ano: already on it
15:12:07 <n0ano> edleafe, tnx
15:12:39 <n0ano> I think that's all we can say about current patches (keep fingers crossed the the remaining ones get merged)
15:13:06 <n0ano> #topic Vancouver design summit
15:13:28 <n0ano> I wanted us to start thinking about what we want in a scheduler session at Vancouver...
15:13:45 <n0ano> most of the ground work for cleaing up the APIs is done...
15:14:09 <n0ano> hopefully, the actual split will be fairly straight forward but we probably want to talk about that a bit
15:14:28 <bauzas> n0ano: I was aiming to talk about the migration path
15:15:01 <bauzas> n0ano: not talking about the project switch, but rather look at the steps for kicking out Gantt
15:15:19 <n0ano> bauzas, in what way, the goal is that the first gantt implementation will be a replacement for the current scheduler, there should be no migration (other than testing)
15:15:44 <edleafe> n0ano: how about the data?
15:15:46 <bauzas> n0ano: erm, I would like to share your optimism, but we're far from just a code split
15:16:09 <bauzas> edleafe: we actually don't care about data, because scheduler doesn't persist anything
15:16:12 <edleafe> n0ano: we really need to define what data will be owned by gantt, and how it will get there from nova
15:16:26 <edleafe> bauzas: but it will after the split
15:16:35 <bauzas> edleafe: that's post gantt-split
15:16:50 <bauzas> edleafe: the current status is that only compute_nodes is readed by Scheduler
15:16:53 <edleafe> bauzas: yes, which we need to plan before we split
15:17:36 <bauzas> edleafe: so my concern is more about how to share objects with Nova, how to release a client, which kind of shipment will be done for the scheduler, work on CI etc.
15:17:37 <edleafe> bauzas: and make it general enough so that cinder and others can use gantt for their scheduling
15:17:47 <bauzas> edleafe: that's far too early to discuss that
15:18:06 <edleafe> bauzas: too early to implement, sure, but we need to have a direction
15:18:29 <bauzas> edleafe: well, we have a direction
15:18:41 <n0ano> I think I agree with both of you, start the discussion but be aware we're talking futures
15:18:59 <edleafe> n0ano: definitely post-liberty
15:19:01 <bauzas> edleafe: and we even have a target design, lucky we are
15:19:24 <bauzas> edleafe: so honestly, we have to stick with libery things
15:19:31 <bauzas> edleafe: I mean for the Summit
15:19:40 <bauzas> and at least for the Nova talks
15:19:58 <bauzas> we can still do a BoF session for talking about that in between us and other peers
15:20:14 <bauzas> we can even take a pod
15:20:22 <n0ano> bauzas, I'd like to at least get people thinking about post-Liberty things, longer term planning can be helpful
15:20:28 <bauzas> but Nova sessions won't be post-L focused
15:20:31 <edleafe> bauzas: my only concern is that splitting becomes the final goal instead of a big first step
15:21:01 <bauzas> n0ano: 100% agreeing with you, but not during Nova sessions, that's it :)
15:21:12 <n0ano> edleafe, splitting was never my end goal, getting other projects to use gantt, that's closer to my end game
15:21:25 <bauzas> n0ano: so, again, pods, BoFs, teasing at the end of a certain presentation
15:21:33 <n0ano> bauzas, +1
15:21:48 <bauzas> I think we all are in violent agreement then
15:21:49 <n0ano> especially the tease at the end of a discussion
15:22:11 <bauzas> let's keep Nova talks for Nova guys, and meet us together if we want to discuss post-L
15:22:43 <n0ano> I don't want to to insular, it would be nice if we can get other projects involved in the post-L ideas
15:22:51 <bauzas> n0ano: hence the BoF thing
15:23:07 <edleafe> agreed - BoF is best for these talks
15:23:19 <n0ano> bauzas, yeah, I'm think of some sort of BoF for cross-project issues
15:23:23 <bauzas> n0ano: we seriously need Cinder peers
15:23:38 <n0ano> bauzas, +2
15:23:43 <bauzas> n0ano: anyway, I don't even have the Design Summit agenda
15:24:03 <bauzas> so I don't know if we'll still rule out cross-project sessions
15:24:22 <n0ano> that's why I wanted to start thinking about things so we can be ready as soon as the schedule becomes visible
15:24:23 <bauzas> but I can talk to Thierry this week
15:24:50 <alex_xu> n0ano: +1
15:25:00 <bauzas> n0ano: hence my point : let's keep focused on Nova split for L, and open a BoF and potentially a cross-project session for the rest
15:25:08 <n0ano> bauzas, if you want to ask him specifically if there are any x-project sessions that'd be great
15:25:14 <bauzas> n0ano: I can
15:25:22 <bauzas> n0ano: he'll be in Grenoble
15:25:39 <n0ano> you can get him in person?
15:25:52 <bauzas> n0ano: lol
15:26:13 * n0ano needs to brush up on France geography :-)
15:26:19 <bauzas> n0ano: I wish I would be as famous as Kanye West, but at least I know Thierry :)
15:27:11 <bauzas> n0ano: you'll have some France maps by the end of the day - due to the plane crash
15:27:33 <n0ano> Ick, too true :-(
15:27:54 <n0ano> I think we're in agreement, let's think on this subject and talk again next week
15:28:11 <n0ano> #topic opens
15:28:17 <bauzas> n0ano: I think we should be ready for opening up an etherpad next week then :D
15:28:18 <n0ano> anyone have anything new for today?
15:28:21 <alex_xu> I need a help on a bug
15:28:27 <bauzas> n0ano: should we make an announcement ?
15:28:27 <alex_xu> #link https://bugs.launchpad.net/nova/+bug/1420102
15:28:29 <openstack> Launchpad bug 1420102 in OpenStack Compute (nova) "Periodic task for quota reservations expire should move out from scheduler" [Wishlist,In progress] - Assigned to Alex Xu (xuhj)
15:28:39 <n0ano> bauzas, sounds good to me
15:28:47 <bauzas> first, alex_xu
15:29:06 <bauzas> alex_xu: good to see you, in particular so late for you
15:29:17 <bauzas> alex_xu: I saw your patch, and I followed the reviews
15:29:19 <alex_xu> The early fix just move the expire periodic task into the conductor, but get some -1. Then I changed another way, but Joe said that way is too complex for now, we are in RC
15:29:30 <alex_xu> bauzas: thanks :)
15:29:35 <bauzas> alex_xu: but IIRC it was about having this done post-K3 right ?
15:29:49 <alex_xu> so the question is whether we want to target that bug for the K?
15:30:00 <bauzas> alex_xu: I think we can defer to L
15:30:08 <edleafe> alex_xu: is it really a bug?
15:30:18 <edleafe> alex_xu: or just a step to separating scheduler?
15:30:20 <alex_xu> edleafe: emm...not a bug
15:30:22 <bauzas> alex_xu: edleafe: it was a wishlist bug, so technically not a bug
15:30:31 <n0ano> alex_xu, at this stage I'd prefer to defer it, it isn't an issue until the split which is definitely after K
15:30:36 <edleafe> yeah, then is should be in L
15:30:52 <alex_xu> ok, cool, we get agreement :)
15:31:14 <alex_xu> I'm ok with L
15:31:32 <bauzas> cool
15:31:39 <bauzas> time for an announcement then
15:31:40 <bauzas> https://review.openstack.org/#/c/160506/8/nova/api/openstack/compute/contrib/hypervisors.py,cm
15:31:44 <bauzas> oops
15:31:46 <bauzas> lol
15:31:51 <bauzas> http://openstacksummitmay2015vancouver.sched.org/event/4ebfd98c565aaf9480d8389344ece74b#.VRGDTsuvhiU
15:32:28 <bauzas> so, please be sure to arrive in Vancouver not later than Monday :)
15:32:28 <n0ano> hmm, wonder why I don't have a picture, I should do something about that
15:32:46 <bauzas> n0ano: something messed up with your profile, because it was duplicated
15:33:22 <n0ano> it's just my attempt at being anonymous :-)
15:33:39 <lxsli> nice, see you then :)
15:33:55 <bauzas> n0ano: I should ask the Daft Punk
15:34:02 <alex_xu> I will be there if I can go, if not I will watch the video also
15:34:14 <edleafe> lxsli: do you want to bring the ripe fruit to throw, or should I? :-P
15:34:32 <n0ano> edleafe, tnx for the warning, I'll bring a shield :-)
15:34:45 <edleafe> n0ano: jsut stand behind bauzas
15:34:51 <bauzas> edleafe: I will prepare myself by running a one-man-show standup before
15:34:55 <n0ano> edleafe, +1
15:35:20 <n0ano> we're in silly time, anything else serious?
15:35:36 <edleafe> nope
15:35:56 <bauzas> back to serious work then, enough chats
15:36:19 <n0ano> most of my serious work `is` chatting :-(
15:36:29 <n0ano> Ok, tnx everyone, talk to you next week (if not before)
15:36:34 <n0ano> #endmeeting