14:00:28 <n0ano> #startmeeting nova-scheduler
14:00:34 <openstack> Meeting started Mon Aug 24 14:00:28 2015 UTC and is due to finish in 60 minutes.  The chair is n0ano. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:38 <openstack> The meeting name has been set to 'nova_scheduler'
14:00:50 <edleafe> o/
14:00:53 <n0ano> anyone here to talk about the scheduler?
14:02:15 * johnthetubaguy lurks
14:02:20 <n0ano> bauzas, was on #nova, I don't see jay though
14:03:25 <bauzas> \o
14:03:38 * bauzas is still a bit jetlagged
14:03:47 <n0ano> lxsli, is the only other regular missing
14:03:57 <n0ano> let's get started anyway
14:04:10 <n0ano> #topic liberty patches
14:04:25 <bauzas> sooo
14:04:35 <n0ano> bauzas, I see you got jenkins happy with your API patch, are you all done now :-)
14:04:49 <bauzas> API patch ? :)
14:04:56 <bauzas> n0ano: you mean https://review.openstack.org/145528 ?
14:05:11 <n0ano> https://review.openstack.org/199205
14:05:47 <bauzas> n0ano: ah that
14:06:26 <bauzas> n0ano: well, the whole chain is done - some stuff still needs to be done on the conductor side, but all the interfaces should ideally be merged by Lib
14:07:00 <n0ano> that's very good, now all we need is to review everything
14:07:11 <bauzas> n0ano: I mean that some helpers can go to hell, but that can be done in Mitaka, since all the APIs would be now using a Spec object
14:07:22 <bauzas> n0ano: yeah, that's the plan
14:07:40 <bauzas> the first one was good enough for alaski and dansmith :)
14:07:53 <n0ano> excellent
14:08:05 <bauzas> I also have a 2nd series https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/allocation-ratio-to-resource-tracker,n,z
14:08:22 <bauzas> which is nearly done, just needing the RT update to DB
14:08:34 <bauzas> and the Scheduler addition
14:08:39 <n0ano> is that on the priorities page, I don't see it
14:08:45 <bauzas> n0ano: nope, not yet
14:08:47 <bauzas> oh
14:08:48 <bauzas> sec
14:09:13 <bauzas> n0ano: it is
14:09:21 <bauzas> https://etherpad.openstack.org/p/liberty-nova-priorities-tracking L57
14:10:02 <n0ano> ah, I see it now, so is that your top priority patch series now?
14:10:15 <bauzas> my top prio is the Spec objectification
14:10:25 <bauzas> that's my 2nd top prio :)
14:10:48 <bauzas> I mean, it's not really useful for cleaning the APIs, but that's also good enough for the scheduler
14:11:12 <bauzas> so, my status here, ready for the review limbo dance of rebasing
14:11:39 <bauzas> the priority going to the Spec stuff, but I'd be happy to see https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/allocation-ratio-to-resource-tracker,n,z merged too
14:11:41 <n0ano> that's actually very cool, frustrating for you but at least the implementation is done
14:12:00 <bauzas> n0ano: well, I'm not frustrated, I have very good feedback
14:12:44 <bauzas> n0ano: it just takes time to agree on some particular bits, in particular when you review a NovaObject because of the fields
14:12:46 <n0ano> I would imagine the rebase feedback loop could get tiresome but, it you're happy with that, great :-)
14:13:01 <bauzas> n0ano: well, it's a matter of getting things done
14:13:15 <bauzas> n0ano: changing a field or removing it is just painful when the code is merged
14:13:28 <bauzas> n0ano: because you can't just pretend it never existed
14:13:39 <n0ano> indeed
14:14:00 <bauzas> so we had some good talks about using a Flavor object or not
14:14:06 <bauzas> that was necessary
14:14:17 <bauzas> (I mean the convo was necessary)
14:14:53 <n0ano> who's we and did you reach a conclusion?
14:14:59 <bauzas> and the consensus was to leave the Flavor object because of silly filters using it, and then work later on changing those filters to not keep that Flavor object which is painful
14:15:21 <bauzas> who: alaski, dansmith and I
14:15:29 <edleafe> n0ano: we want to eventually limit flavors to the API layer only
14:15:58 <bauzas> n0ano: I was btw. a bit surprised to see a spec of you, made in June :p
14:15:59 <n0ano> I wanted to have a session at Mitaka about that but it looks like maybe we've resolved the issue already
14:16:26 <n0ano> bauzas, it got delayed and i was intending to retry it for Mitaka
14:16:32 <alaski> I don't want a flavor on the spec object, but for now it's mostly an implementation detail and the individual components are exposed on the object as well
14:16:53 <bauzas> so, the most boring stuff was https://github.com/openstack/nova/blob/master/nova/scheduler/filters/type_filter.py#L35
14:17:40 <bauzas> or even https://github.com/openstack/nova/blob/master/nova/scheduler/filters/type_filter.py#L58
14:19:11 <n0ano> that's an interesting filter, do people really use that?
14:20:07 <bauzas> n0ano: well, that's a good question, but since it's shipped, we have to keep at least a compatibility
14:20:23 <bauzas> I mean a compatible behaviour
14:20:34 <n0ano> if every there was a filter ripe for deprecation I think that would be it
14:20:55 <bauzas> yeah, we haven't yet deprecated a filter
14:21:05 <bauzas> and that would be an interesting talk
14:21:13 <bauzas> I have 3 of those in my target
14:21:27 <bauzas> ComputeFilter, TrustedFilter and those flavor filters
14:21:49 <bauzas> but those flavor filters are by far less silly than the 2 above I mentioned
14:22:23 <n0ano> potentially an interesting session in Tokyo
14:22:29 <bauzas> so, we can leave a room for a discussion about any deprecation rules in Mitaka, would be glad to kick that off
14:22:56 <bauzas> well, maybe not a session, but surely a bullet point in the sched area
14:23:05 <n0ano> bauzas, +1
14:23:20 <bauzas> moving on ?
14:23:25 <n0ano> WFM
14:23:33 <n0ano> anything else on patches?
14:23:38 <bauzas> I don't have any status for jay
14:23:53 <bauzas> I know that he had made good stuff with the pci broken things
14:23:59 <bauzas> I'm mostly reviewing
14:24:08 <n0ano> sounds good
14:24:09 <bauzas> https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:pci-cleanup,n,z
14:24:22 <n0ano> in that case
14:24:27 <n0ano> #topic opens
14:24:34 <n0ano> anything new for today?
14:24:39 <bauzas> which is a prereq for https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:resource-objects,n,z
14:25:01 <bauzas> so, yeah, folks have plenty of changes to reviews
14:25:46 <bauzas> so, opens ?
14:25:59 <n0ano> that's the topic, not hearing anything
14:26:25 <edleafe> nothing from me
14:26:55 <n0ano> OK, then time to do our reviews (unless you're me which means it's time for my next meeting :-(
14:27:13 <n0ano> tnx everyone, we'll talk next week.
14:27:13 <bauzas> fair enough
14:27:22 <n0ano> #endmeeting