14:00:12 <edleafe> #startmeeting nova_scheduler
14:00:13 <openstack> Meeting started Mon Aug 14 14:00:12 2017 UTC and is due to finish in 60 minutes.  The chair is edleafe. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:16 <openstack> The meeting name has been set to 'nova_scheduler'
14:00:27 <edleafe> Good UGT morning! Who's here?
14:00:57 <edleafe> With RC1 cut, and many cores on holiday, this should be a short meeting
14:01:29 <dtantsur> o/
14:02:28 <dtantsur> a room full of people :D
14:03:28 <edleafe> Well, let's give people a few minutes to join. Otherwise, we can just wrap this up and get back to work
14:05:24 * cdent wanders in
14:06:01 <edleafe> Well, instead of going through the list of things, let's just do:
14:06:07 <edleafe> #topic Open Discussion
14:06:19 <edleafe> If you have anything to discuss, now's the time
14:06:29 <dtantsur> I think this is the last bit to fix ironic with custom resource classes: https://review.openstack.org/492964
14:06:40 <cdent> I’ve got something I’d like some eyes on, to see if the approach is suitable:
14:06:52 <cdent> https://review.openstack.org/#/c/493536/
14:07:31 <edleafe> dtantsur: I have that patch on my review list for today
14:07:37 <dtantsur> cool
14:07:39 <cdent> I’m also curious if anybody has some ideas on strategy to find and tie up all our loose ends this week.
14:07:52 <cdent> I’m not sure if we have many or not.
14:08:02 <edleafe> cdent: well, first step would be: identify what the loose ends are
14:08:13 <cdent> yes
14:08:18 <cdent> <- dunno
14:08:31 <edleafe> Me neither
14:08:43 <cdent> but I’ll try to dig my way out somehow
14:08:52 <edleafe> As dtantsur pointed out, https://review.openstack.org/487954 merged and I'll propose a backport to stable/pike
14:09:29 <edleafe> cdent: should the ksa patch also be in Pike?
14:09:44 <edleafe> dtantsur: same question about your patch
14:09:58 <dtantsur> yep
14:10:21 <cdent> I’m not certain. It doesn’t count as a regression: we’ve had that problem since the dawn of placement in the resource tracker
14:10:45 <cdent> our advice has always been: get placement going before your restart/upgrade your compute nodes
14:10:58 <cdent> but I guess in the world of asynchronous config mgt that’s not always happening
14:11:12 <edleafe> cdent: understood
14:11:30 <cdent> I we are good people who want to make people’s lives nice we would include it, but I don’t think our policy accounts for “good” and "nice".
14:11:38 <edleafe> is there any potential risk in doing this? I don't see any, but I'm not ksa expert
14:12:30 <cdent> the cost/risk is that we’ll being hitting keystone for a new service catalog more often until placement is up
14:12:46 <cdent> since we _require_ placement in pike (and earlier) I don’t see this as an issue
14:13:04 <edleafe> So if they do things the advertised way, there should be no impact
14:13:16 <cdent> correct
14:13:24 <edleafe> If they do things out of order, there may be a slight rise in the number of keystone calls
14:14:08 <cdent> correct
14:14:56 <edleafe> Well, I suppose it's all up to dansmith at this point. Tests will be added soon?
14:15:01 <cdent> If they hit the bug, for whatever reason, they will have confused n-cpu process for however long the caching of the service catalog last
14:15:20 <cdent> yes, basically my next thing to do after this meeting
14:16:00 <edleafe> cdent: cool. Will review then
14:16:05 <edleafe> Anything else today?
14:16:52 <edleafe> going once...
14:17:02 <edleafe> going twice...
14:17:11 <edleafe> #endmeeting