15:00:03 <n0ano> #startmeeting gantt
15:00:04 <openstack> Meeting started Tue Dec  9 15:00:03 2014 UTC and is due to finish in 60 minutes.  The chair is n0ano. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:07 <bauzas> \o/
15:00:08 <openstack> The meeting name has been set to 'gantt'
15:00:14 <n0ano> anyone here to talk about the scheduler
15:00:27 <bauzas> \o
15:00:33 <alex_xu> hello, all, first time join this meeting :)
15:00:44 <n0ano> alex_xu, welcome
15:00:49 <edleafe> \o
15:00:53 <alex_xu> n0ano, thanks! :)
15:00:54 <bauzas> alex_xu: welcome on board
15:01:07 <alex_xu> bauzas, thanks!
15:01:15 <bauzas> alex_xu: isn't quite an awful time for you ?
15:01:43 <alex_xu> bauzas, it's fine, just late sleep time about one hour
15:01:54 <bauzas> alex_xu: GMT+8 ?
15:02:12 <alex_xu> bauzas, yes, now it's 11pm
15:02:16 <bauzas> ah
15:02:18 <n0ano> bauzas, should be 11PM his time (I don't even go to bed then)
15:02:39 <n0ano> wow, my timezone is correct, that's a first :-)
15:02:42 <n0ano> anyway, let
15:02:47 <n0ano> anyway, let's start
15:02:48 <bauzas> n0ano: hell, I know how it's cool to attend 11pm's meetings, so I truly appreciate his presence :)
15:03:24 <n0ano> #topic forklift status
15:03:39 <bauzas> I guess it's the only topic we have to cover ?
15:03:42 <n0ano> bauzas, I see you got another spec approved, congratulation
15:03:50 <n0ano> bauzas, that & opens
15:03:52 <bauzas> n0ano: yeah that's cool
15:04:01 <bauzas> n0ano: the one dependent is having one +2
15:04:18 <bauzas> maybe some pointers could be useful actually
15:04:29 <bauzas> #link https://wiki.openstack.org/wiki/Gantt/kilo#Tasks
15:05:02 <bauzas> #info https://review.openstack.org/#/c/127610/ has been approved
15:05:27 <bauzas> https://review.openstack.org/#/c/127612/ is having one +2
15:05:38 <bauzas> so we're in a good shape for both acceptance by K2
15:06:01 <n0ano> bauzas, 127612 is rather critical, do you see a problem with getting the second +2
15:06:04 <bauzas> even if we can't make the second for spec freeze, I think we have good chances to get an exception
15:06:33 <bauzas> n0ano: nope, not really, maybe the filt_props merge with req_spec could be debatable
15:06:41 <bauzas> n0ano: but nothing really needing a -2
15:07:04 <n0ano> sounds cautiously optimistic then
15:07:10 <bauzas> n0ano: eh
15:07:23 <n0ano> I prefer to let that work it's way through
15:07:23 <bauzas> n0ano: sounds pragmaticly optimistic :)
15:07:32 <n0ano> about the others
15:07:46 <n0ano> #link https://review.openstack.org/#/c/127609/
15:07:49 <bauzas> n0ano: https://review.openstack.org/#/c/89893/ is having a -2 from johnthetubaguy
15:08:06 <bauzas> jaypipes: around ?
15:08:11 <bauzas> jaypipes: Gantt meeting eh ;)
15:08:38 <bauzas> n0ano: I can see some implementation proposal which seems good to me
15:08:46 <bauzas> n0ano: this spec is having -1 tho
15:09:05 <bauzas> n0ano: but it seems it's a nitpikc
15:09:11 <bauzas> n0ano: so I'm still confident
15:09:21 <n0ano> we still on 89893, the only thing I see is an x from garbutt
15:09:25 <bauzas> n0ano: if you see the -1, it comes from something missing
15:09:33 <bauzas> n0ano: yeah, so 89893
15:09:51 <bauzas> n0ano: so, I had a crossbar from him because he was not liking my former proposal
15:10:12 <bauzas> n0ano: so we google hangouted 2 weeks ago
15:10:29 <n0ano> yeah, his comment was like the approach, issues with the implementation, those should be resolvable
15:10:31 <bauzas> n0ano: and the new PS is based on our discussions
15:10:54 <bauzas> n0ano: unfortunately, john is on travel these days, so that's hard to get him
15:11:05 <bauzas> n0ano: in order to ask him to remove his now procedural -2
15:11:27 <n0ano> my concern is the spec freeze next week, are we at risk of missing that?
15:11:42 <bauzas> n0ano: that said, still on the spec, I'm really interested in getting feedback from you and the team
15:11:50 <bauzas> n0ano: I don't think so
15:12:11 <bauzas> n0ano: by discussing with john, I've been told that priorities are in good shape for asking exceptions
15:12:27 <edleafe> And 138444 will follow the same pattern as 89893, so that should be done right after
15:12:27 <bauzas> n0ano: and both missing approvals are quite having consensus now
15:12:43 <bauzas> edleafe: yeah, that sounds feasible
15:12:56 <bauzas> n0ano: to be clear, K1 is not the spec freeze
15:13:05 <n0ano> bauzas, edleafe cool, so the table looks iffy but the reality is we're in reasonable shape
15:13:12 <bauzas> n0ano: that's somewhere in between K1 and K2
15:13:27 <n0ano> bauzas, that's not my understanding, I thought from the Nova meeting last week 12/18 was the spec freeze date
15:13:34 <bauzas> n0ano: but I will chase this up in the next Nova meeting on Thurs
15:13:50 <bauzas> n0ano: I hope to see a spec "proposal" freeze
15:14:09 <n0ano> bauzas, good to confirm that but I think we're in good shape either way
15:14:17 <bauzas> n0ano: as I said, things are different if you're working on a priority thing
15:14:25 <n0ano> bauzas, +1
15:14:25 <bauzas> n0ano: agreed
15:14:46 <bauzas> n0ano: so, in front of implementation patches
15:15:01 <bauzas> n0ano: I have one small tactical victory
15:15:19 <bauzas> that's fresh news
15:15:19 <bauzas> https://review.openstack.org/#/c/126372/19
15:15:21 <bauzas> is merged
15:15:48 <n0ano> I remember that, good news
15:15:56 <bauzas> that's the first patch of the bp/detach-service-from-computenode series
15:16:09 <bauzas> the next 3 ones are also in a good shape
15:16:14 <bauzas> *but*
15:16:43 <bauzas> somewhere in the series, I'm objectifying calls to compute_nodes table
15:16:59 <bauzas> I can try to delay this patch the farest I can
15:17:16 <bauzas> because that's blocked due to pci_stats missing
15:17:27 <bauzas> in the ComputeNode Object
15:17:42 <n0ano> bauzas, is that blocked on jay's resource object model changes?
15:18:26 <bauzas> n0ano: not at all
15:18:47 <bauzas> n0ano: this patch is independent
15:19:06 <n0ano> well that's good, I worry when both you and jay are talking about objects in different areas
15:19:07 <bauzas> n0ano: basically, I'm changing how scheduler, cells and API are accessing compute_nodes table
15:19:30 <bauzas> n0ano: all the calls will be done using the ComputeNodeList.get_all() object method
15:19:45 <bauzas> jaypipes was also working on the same thing in the HostManager
15:20:01 <n0ano> but you both have to deal with the PCI info
15:20:12 <bauzas> *but* both of us are blocked because we can't yet use the ComputeNode object as it's missing the pci_stats thing
15:20:17 <bauzas> n0ano: exactly
15:20:18 <bauzas> https://review.openstack.org/#/c/137847/
15:20:28 <bauzas> n0ano: ^ is the patch fixing that
15:20:39 <bauzas> n0ano: led by Paul Murray
15:20:51 <bauzas> n0ano: but some details need to be fixed
15:21:23 <bauzas> n0ano: so as I'm blocked, I'm leaving this bp/detach-service-from-computenode series
15:21:23 <n0ano> I trust paul but it looks like he's got some issues to resolve, might take  a little work
15:21:31 <bauzas> n0ano: exactly
15:22:10 <bauzas> n0ano: so my strategy is to descope from this blueprint (which was targeted for K1) and work on the request-spec-object BP which was targeted for K2 - in advance
15:22:12 <jaypipes> hey guys, sorry... stepped away for a bit.
15:22:36 <bauzas> n0ano: and leave the good bits from detach-service to be merged when possible - until the object patch
15:22:50 <n0ano> bauzas, sounds like a plan
15:23:14 <bauzas> n0ano: so I actually provided a first patch for request-spec-object
15:23:25 <bauzas> https://review.openstack.org/139684
15:23:43 <bauzas> that's not that important but it was left in the spec to upgrade the RPC API version
15:23:46 <jaypipes> yeah, guys, I was at an offsit all last part of last week so still playing catchup on reviews and emails. :(
15:23:48 <bauzas> so did I
15:24:05 <bauzas> jaypipes: no worries, feel free to rewind the backlog ;)
15:24:08 <n0ano> jaypipes, no excuses, the internet is everywhere :-)
15:24:32 * bauzas can see the difference of hospitality in between bauzas and n0ano
15:24:49 * n0ano has a tendency to crack the whip a bit :-)
15:25:11 <bauzas> :)
15:25:20 <n0ano> jaypipes, summary, yet another spec approved, cautiosly optimistic, lots of reviews needed
15:25:35 <bauzas> eh, we're working ;)
15:25:52 <n0ano> s/cautiosly/cautiously
15:26:32 <n0ano> jaypipes, anyway, the big one for you is https://blueprints.launchpad.net/nova/+spec/resource-objects any update on that
15:27:24 <jaypipes> n0ano: waiting on follow up round of reviews from folks, including danpb. But it's generally waiting for the work from ptm and bauzas, so I've been focusing on reviewing those patch series.
15:27:34 <jaypipes> along with all the nUMA objects stuff, which was a crap-ton of work.
15:28:17 <n0ano> sounds like there's no major problem, `just do it`
15:28:55 <bauzas> jaypipes: are you blocked on me ?
15:29:00 <bauzas> jaypipes: orly ?
15:29:11 <bauzas> jaypipes: how can I help you then ?
15:29:57 <jaypipes> bauzas: nope, not blocked on you... just need to wait for your objects changes to land before I continue with more work on the resourc eobjects.
15:31:12 <bauzas> jaypipes: you mean the RequestSpec object ?
15:31:36 <bauzas> jaypipes: or the objectification of calls to compute_nodes table ?
15:31:51 * bauzas thinks the word "object" is definitely too wide now :)
15:32:16 <n0ano> bauzas, not too wide, we just have a lot of work related to objects
15:32:31 <jaypipes> bauzas: the objectification of calls to the compute_nodes table.
15:32:36 <bauzas> jaypipes: oh ok
15:33:07 <bauzas> jaypipes: as I said previously, I'm not targeting to work on this until paul's patch merges
15:33:16 <jaypipes> bauzas: yes, understood.
15:33:23 <bauzas> jaypipes: so https://review.openstack.org/137847 is the top priority then :)
15:33:28 <jaypipes> yes
15:33:32 <bauzas> n0ano: ^
15:33:50 <n0ano> sigh, xchat just crashed, I'm back now, what did I miss?
15:34:01 <bauzas> we were just asking you to work
15:34:03 <bauzas> :)
15:34:14 <bauzas> n0ano: https://review.openstack.org/137847 is top prio
15:34:19 <n0ano> you can ask, whether you can get me to do it is a different thing
15:34:22 <bauzas> n0ano: so reviews are welcome
15:34:37 <n0ano> OK, good to know, as always - review, review, review
15:35:01 <bauzas> n0ano: I'll ask paul if he has time for fixing the comments before I'm putting this patch on the magical etherpad
15:35:39 <bauzas> the magical etherpad being the one figuring all the patches related to the priorities
15:35:48 <n0ano> bauzas, OK, let me know how that turns out, I don't want all of us beating up on paul, one is enough
15:35:56 <bauzas> the ones we want to have core coverage
15:36:22 <bauzas> n0ano: eh he's English, I'm French, I'm fine if we're beating him :)
15:36:41 <n0ano> bauzas, +1 (good one)
15:37:07 <n0ano> I think I'll add a footnote to our task table point out this dependency, it's important
15:37:17 <jaypipes> bauzas: I'm British too :)
15:37:25 <bauzas> jaypipes: we all do errors
15:37:31 <jaypipes> ha!
15:37:55 <n0ano> hah, my ancestry is English/German, bauzas `really` hates me :-)
15:37:58 <bauzas> that's really sad to talk about paul and not even lt him
15:38:05 <jaypipes> ok, time to wrap up this meeting :)
15:38:17 <n0ano> jaypipes, not quite...
15:38:23 <n0ano> #topic opens
15:38:33 <n0ano> anyone have anything new for today?
15:39:15 * n0ano listens to the crickets
15:39:25 <bauzas> eh
15:39:33 <bauzas> we're in winter
15:39:37 <alex_xu> I'm try to add rescheduling and fix race of instance group https://review.openstack.org/139843, welcome review
15:39:43 <n0ano> OK, we can close but remember - reviews
15:39:58 <bauzas> alex_xu: yeah I reviewed your series
15:40:01 <jaypipes> alex_xu: will do. thx for letting us nkow!
15:40:17 <alex_xu> bauzas, thanks for your comment!
15:40:20 <alex_xu> jaypipes, thanks
15:40:25 <n0ano> OK, tnx everyone, talk next week
15:40:28 <n0ano> #endmeeting