14:00:13 <edleafe> #startmeeting nova_scheduler
14:00:14 <openstack> Meeting started Mon Jun 13 14:00:13 2016 UTC and is due to finish in 60 minutes.  The chair is edleafe. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:17 <openstack> The meeting name has been set to 'nova_scheduler'
14:00:21 <edleafe> who's around?
14:00:25 <mlavalle> o/
14:00:26 <takashin> o/
14:01:19 <bauzas> \o
14:01:25 <cdent> o/
14:01:53 <edleafe> small crowd - lots of empty seats!
14:02:15 <edleafe> wait another minute for latecomers to arrive
14:02:18 <jaypipes> morning.
14:02:34 <diga> o/
14:03:18 <edleafe> Nobody added anything to the agenda, so I'm guessing that this will be quick
14:03:23 <edleafe> #topic specs
14:03:41 <edleafe> Anyone have a spec they'd like to focus on?
14:04:09 <jaypipes> edleafe: I'm still working on redoing the resource-providers-allocations and dynamic-resource-classes specs. been focusing on trying to get cdent's patches around inventories reviewed.
14:04:36 <_gryf> o/
14:04:41 <woodster_> o/
14:04:54 <edleafe> jaypipes: cool. Any estimate of when the reworked specs will be up for review?
14:05:31 <bauzas> FWIW, my current review effort is on https://review.openstack.org/#/c/326440/14/
14:05:56 <jaypipes> edleafe: by next monday.
14:06:00 <edleafe> bauzas: yes, that's the first of cdent's series
14:06:12 <jaypipes> edleafe: out in CA monday through wednesday this week...
14:06:48 <cdent> bauzas, edleafe, jaypipes I just pushed a changes across that stack
14:06:53 <edleafe> #info jaypipes to be useless for most of this week :)
14:06:58 <edleafe> hehehe
14:07:03 <edleafe> #undo
14:07:04 <openstack> Removing item from minutes: <ircmeeting.items.Info object at 0x7fe237670d90>
14:07:04 <cdent> mostly in the api and the delete_inventory stuff (added add and update)
14:07:06 <jaypipes> cdent: coolio. will try to re-review while in meetings this morning.
14:07:08 <bauzas> cdent: yeah just saw
14:07:14 <jaypipes> edleafe: :)
14:07:26 <jaypipes> cdent: I really like the first effort on the API stuff.
14:07:33 <cdent> awesome, thanks
14:07:34 <jaypipes> cdent: ++ on that.
14:07:45 <jaypipes> cdent: as little usage of nova.wsgi as possible.
14:07:46 <edleafe> yes, that's looking good, cdent
14:07:53 <cdent> jaypipes: I got rid of the rest of it
14:07:58 <jaypipes> sweetness.
14:08:09 <edleafe> #topic reviews
14:08:20 <edleafe> we've been kind of on this topic anyways...
14:08:23 <jaypipes> edleafe: and I know I owe you more review and discussion time on the qualitative side of the equation.
14:08:37 <bauzas> jaypipes: cdent: not sure I like https://review.openstack.org/#/c/293104/57/nova/api/openstack/placement/handlers/aggregate.py@40 but let's discuss this in the change itself
14:08:55 <edleafe> jaypipes: thanks. I know that won't make it for newton, but I'd appreciate calrifying the direction
14:09:04 <jaypipes> bauzas: that is updating the aggregates for a resource provider.
14:09:17 <bauzas> jaypipes: yeah, I think it's just a naming problem
14:09:20 <jaypipes> bauzas: not updating the aggregate info itself.
14:09:31 <jaypipes> bauzas: sure fair enough.
14:09:33 <bauzas> jaypipes: yup, fortunatly
14:09:54 <cdent> bauzas: as I say in the commit message aggregates and allocations are currently broken, so that stuff will change quite a bit
14:10:12 <bauzas> okay
14:10:13 <jaypipes> BTW, best thing anyone can do right now for reviews is get the f-ing Citrix XenServer CI out of the gate rotation. It's crapped out yet again.
14:10:30 <cdent> I'm going to break the api patch into multiple commits, either later today or tomorrow
14:10:56 <jaypipes> cdent: ++ the sooner the better. I offered a possible patch split strategy in my review yesterday.
14:11:13 <cdent> ✔
14:12:11 <edleafe> So it looks like cdent patch series beginning with https://review.openstack.org/#/c/326440 should get as many eyes as possible
14:12:45 <bauzas> cdent: jaypipes: I'm a bit concerned by seeing the API handlers quite different from what I'm used to, but I guess it's something you wanted
14:13:00 <cdent> bauzas: it wasn't just us
14:13:10 <cdent> Multiple people requested that it not be like nova.wsgi
14:13:35 <jaypipes> bauzas: yes.
14:14:00 <bauzas> which means that I need to learn another foreign language when reviewing, and other people to speak another language, but okay
14:14:23 <jaypipes> bauzas: pls talk with cdent and sdague about that API stuff in #openstack-nova
14:14:35 <bauzas> FWIW, I'd by far wait for other reviewers giving their own opinions before we merge that one
14:14:56 <jaypipes> bauzas: pls discuss that with sdague :)
14:15:15 <jaypipes> bauzas: not trying to shush you off, just need to leave for a meeting shortly ;)
14:15:34 <bauzas> jaypipes: I'm just saying I'm not against, just that it needs a bit of people saying +1/+2 before we merge it :)
14:15:53 <bauzas> but okay
14:16:02 <edleafe> bauzas: good idea. Let's keep the discussion going on the patches themselves
14:17:01 <edleafe> There is also the multiple host spec: https://review.openstack.org/276840
14:17:29 <bauzas> anyway, I'm not an API expert, so that's why I'd love some other people hitting that change :)
14:17:46 <edleafe> I've posted a WIP patchset as a strawman for discussing implementation, too: https://review.openstack.org/327809
14:17:52 <edleafe> Feedback is welcome
14:18:14 <bauzas> edleafe: well, the problem is that check-destinations was not a prio
14:18:35 <bauzas> edleafe: that leaves me thinking that multiple-hosts is not, too
14:18:51 <jaypipes> edleafe: I read through that and didn't have anything negative on it. ++ from me. I'll add a review.
14:18:52 <bauzas> but let's discuss that with the other folks (again)
14:19:05 <edleafe> bauzas: hmmm... that wasn't my impression, but I could be wrong
14:19:27 <edleafe> jaypipes: thx
14:19:30 <bauzas> edleafe: no worries, just explaining you that my own BP was "medium" :)
14:19:41 <edleafe> bauzas: ack
14:20:16 <edleafe> OK, any other reviews that need some love?
14:21:03 <edleafe> #topic Opens
14:21:12 <diga> jaypipes, edleafe : is there anything I can contribute. I am eager to start on nova scheduler
14:21:36 <edleafe> diga: how familiar are you with the current code?
14:21:54 <diga> I know about nova code
14:22:56 <diga> edleafe: I have written one filter scheduler at havana release - https://blueprints.launchpad.net/nova/+spec/license-cost-aware-vm-placement
14:22:59 <diga> for this BP
14:23:08 <diga> but somehow it is not approved
14:23:17 <edleafe> diga: I would start reviewing the changes for the Resource Providers specs. One is: https://blueprints.launchpad.net/nova/+spec/resource-providers-allocations
14:23:37 <diga> edleafe: okay
14:24:10 <edleafe> diga: the reason I suggest those specs is that we're changing how resources are represented for the scheduler
14:24:25 <diga> edleafe: okay
14:24:42 <edleafe> diga: Once you get the basic design that we're creating, the code to implement it will make a lot more sense
14:24:49 <diga> edleafe: I will go through this
14:24:53 <diga> edleafe: sure
14:25:00 <diga> +1
14:25:10 <edleafe> diga: great! We can use all the help we can get! :)
14:25:31 <diga> edleafe: Thanks for considering :)
14:25:39 <bauzas> well, not sure the resource-providers blueprints are the easiest to review, but okay...
14:25:56 <edleafe> bauzas: oh, yes, not the easiest at all
14:26:01 <jaypipes> sorry folks, I need to run off now...
14:26:22 <edleafe> bauzas: but if diga already has some familiarity with the scheduler, that's the bigggest change in direction since Havana
14:26:38 <edleafe> jaypipes: happy travels!
14:26:58 <bauzas> edleafe: yup, at least saying that's worth knowing what's going to be changed
14:27:04 <bauzas> ie. reading the specs
14:27:15 <edleafe> bauzas: 'zactly
14:28:02 <edleafe> Anything else to discuss? Or can I give you all back 32 minutes of your day?
14:28:26 <bauzas> <3
14:28:33 <edleafe> #endmeeting