14:00:23 <n0ano> #startmeeting nova-scheduler
14:00:24 <openstack> Meeting started Mon Feb  8 14:00:23 2016 UTC and is due to finish in 60 minutes.  The chair is n0ano. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:27 <openstack> The meeting name has been set to 'nova_scheduler'
14:00:32 <n0ano> anyone here to talk about the scheduler?
14:00:32 <cdent> o/
14:00:43 <_gryf> \o
14:01:15 * cdent was hoping for a dance chain
14:01:15 <edleafe> \o
14:01:25 <Yingxin> o/ happy chinese new year
14:01:34 <edleafe> cdent: too early on a MOnday for that!
14:01:47 * cdent dances anway
14:01:50 <cdent> anyway!
14:01:58 <n0ano> who cares about new years, local team won the super bowl :-)
14:02:18 <edleafe> Yingxin: happy chinese new year to you! Why are you at work? ")
14:02:48 <jaypipes> o/
14:03:14 <Yingxin> edleafe: it is always not my work time :P
14:03:41 * bauzas waves
14:03:48 <n0ano> looks like we have approximate quorum, let's start
14:04:01 <n0ano> #topic  Patches/Reviews – https://etherpad.openstack.org/p/mitaka-nova-priorities-tracking
14:04:41 <n0ano> according to the tracking page we only have 4 reviews outstanding, is that correct or are we missing some from the page?
14:04:48 * cdent adds a couple things
14:04:56 * n0ano says as more are being added realtime
14:05:10 <bauzas> n0ano: I'm about to add https://review.openstack.org/#/c/273104/2
14:05:15 <cdent> and I have another one in progress but not submitted yet (all resource-provider related)
14:05:30 <bauzas> n0ano: once I fix my stupid miss
14:06:14 <n0ano> jaypipes, I was going to point out the resource provider, looks like those will be added soon
14:06:29 <johnthetubaguy> cdent: how is that generic pools spec coming along?
14:06:37 * cdent looks at jaypipes
14:07:14 <jaypipes> johnthetubaguy: will try to get the next rev up today. sorry for delay.
14:07:36 <johnthetubaguy> just feels very very late for mitaka now, but lets see
14:08:25 <bauzas> we can target to land the respective implementations for the 2 first specs
14:10:21 <johnthetubaguy> bauzas: yep, no need to stop the first two
14:10:28 <johnthetubaguy> they should be all set for mitaka now
14:10:57 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/resource-classes
14:11:06 <johnthetubaguy> #link https://blueprints.launchpad.net/nova/+spec/resource-providers
14:11:16 <johnthetubaguy> I think those are the ones that are good to go right now
14:11:19 <cdent> I expect to have the implementations for those first two in the first half of this week, although I expect there will need some change in review.
14:11:41 <bauzas> +
14:11:45 <bauzas> ++ even
14:11:55 <johnthetubaguy> cdent: sounds good
14:11:57 <cdent> Although I'm going to need to bother someone for a bit of advice on some of the details about versioned objects which are composed of others
14:12:02 <n0ano> cdent, that'd be great, just be sure to update the etherpad appropriately
14:12:15 <cdent> roger that, n0ano
14:12:18 <johnthetubaguy> cdent: you mean the object fields in an object?
14:12:24 <cdent> johnthetubaguy: yeah
14:13:00 <johnthetubaguy> cdent: cools, can ask questions in channel, thats stuff did get simpler quite recently
14:13:16 * cdent nods okay
14:13:42 <n0ano> are there any specific reviews people want to discuss right now?
14:13:49 <bauzas> n0ano: yup
14:13:52 <bauzas> like I said
14:13:59 <bauzas> https://review.openstack.org/#/c/273104/
14:14:09 <bauzas> given that's also something the live-migrate team wants
14:14:23 <bauzas> well, once I'll provide the respective live-migrate patch
14:14:50 <bauzas> I'd love to get subteam reviews for that one + the ones I'll do for live-migrate and unshelve
14:15:09 <bauzas> which are the 3 conductor methods calling the scheduler by dumbly mocking out a fresh new request spec
14:15:21 <edleafe> bauzas: added to my queue
14:15:37 <johnthetubaguy> bauzas: we totally need to fix up that scheduler_utils stuff at some point, is that in the works already?
14:15:52 <bauzas> johnthetubaguy: that will be the last patch of the series I guess
14:16:10 <johnthetubaguy> bauzas: would it not be cleaner as the first patch?
14:16:29 <bauzas> johnthetubaguy: not really, it would add lots of modifications for something not that clear yet
14:16:45 <johnthetubaguy> bauzas: OK
14:17:38 <bauzas> johnthetubaguy: but we can discuss that in the review for sure
14:17:44 <edleafe> is there a list of scheduler_utils changes that you are looking to make?
14:18:17 <johnthetubaguy> edleafe: basically, return spec_obj is what I am expecting to see, at some point
14:18:26 <bauzas> edleafe: I was thinking of build_request_spec and set_vm_notif
14:18:28 <johnthetubaguy> rather than the legacy dict thing
14:18:42 <bauzas> but since there are lots of callers, that's a pretty hairy change :/
14:18:50 <edleafe> bauzas: yep
14:19:08 <bauzas> hence my point above
14:19:41 <edleafe> ok - I could base it on the last of your series, and aim for integrating in newton
14:19:51 <bauzas> and once all the conductor methods would be using the spec obj, it should be far easier to remove all the conditionals I was about to write eg. in _schedule_destination()
14:20:39 <johnthetubaguy> bauzas: I guess we need to think about bumping the major RPC version just after mitaka-3
14:20:53 <johnthetubaguy> bauzas: so we can do that tidy up next cycle
14:21:37 <bauzas> johnthetubaguy: speaking of the conductor RPC API, sorry ?
14:23:02 <bauzas> (uploading the latest PS by the time I type)
14:23:03 <bauzas> anyway
14:23:19 <bauzas> let's discuss that in the review, I'm all up for comments :)
14:23:35 <n0ano> bauzas, comments in reviews are good :-)
14:23:55 <bauzas> n0ano: moving on ? you probably had hard time yesterday with broncos, so I don't want to make you asleep
14:24:33 <n0ano> bauzas, tnx for your concern, I just went home tobed, no hard partying :-)
14:24:40 <n0ano> anyway
14:24:56 <n0ano> #topic Bugs - https://bugs.launchpad.net/nova/+bugs?field.tag=scheduler
14:25:23 <n0ano> I got commitment for our Texas group to look into sched bugs but doesn't look like any progress yet, hopefully soon
14:25:55 <n0ano> I bug count is sticking at ~38, it would be nice to get that count to decrease as some point
14:26:06 <johnthetubaguy> bauzas: yes, possibly also the scheduler one, would have to go did
14:26:24 <johnthetubaguy> n0ano: is that the OSIC team?
14:27:11 <n0ano> johnthetubaguy, maybe, I don't do acronyms well, it's the joint RackSpace/Intel lab in San Antonio
14:28:28 <n0ano> moving on
14:28:32 <n0ano> #topic opens
14:28:35 <johnthetubaguy> n0ano: yeah, thats them, its just they are focusing on live-migrate bugs right now, although having a good fully triaged list of bugs for folks to work on is no bad thing
14:29:21 <cdent> I've got this non-priority doc-only change that's been pending for a while: https://review.openstack.org/#/c/262561/ would be nice to dispense with it
14:29:39 <n0ano> johnthetubaguy, that's reasonable, I did get a commitment that they would look into sched bugs but that the sched would be a secondary priority, better than nothing
14:29:52 <johnthetubaguy> n0ano : +1
14:30:39 <johnthetubaguy> cdent: I wonder if we do want a patch that tries to talk about the overall picture post mid-cycle?
14:30:53 <johnthetubaguy> cdent: seem like most of the path is defined now
14:31:04 <johnthetubaguy> you know, from 10k feet
14:32:05 <cdent> johnthetubaguy: I think that's a fine idea, but I wasn't sure if/when/who time for that, and I started this one when I was considerably less well informed than I am now (which is not to say that I'm well informed now...). So I figured as a stop gap...
14:32:22 <johnthetubaguy> cdent: thats cool
14:32:29 <bauzas> yeah, providing the epic would be cool
14:32:42 <bauzas> I second that
14:33:19 <johnthetubaguy> I might attempt something, if nothing happens soon, but would love it if someone else gets there first
14:33:20 <n0ano> cdent, trying to write it up would be a great way to become more informed, I'd encourage you to try, we'll all help with reviews
14:33:34 <johnthetubaguy> need it merged well before the summit, ideally
14:33:39 <cdent> n0ano: I'm currently _very_ booked on priority patches...
14:33:52 <n0ano> cdent, understood
14:34:05 <johnthetubaguy> cdent: yeah, the patches are key right now, this is probably for post mikta-3 when all those have to have merged
14:34:17 * cdent nods
14:34:57 <n0ano> anyone have anything else new?
14:35:40 <n0ano> hearing crickets
14:36:00 <n0ano> OK, tnx everyone and keep those reviews coming
14:36:05 <n0ano> #endmeeting