14:00:53 #startmeeting nova_scheduler 14:00:54 Meeting started Mon Mar 13 14:00:53 2017 UTC and is due to finish in 60 minutes. The chair is edleafe. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:56 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:58 The meeting name has been set to 'nova_scheduler' 14:01:04 \o 14:01:05 \o 14:01:06 Good UGT morning! Who's here? 14:01:11 orz 14:01:33 morning :) 14:02:14 With the recent DST shift, I now have time to caffeinate before these meetings :) 14:02:28 * cdent holds on to his hat 14:02:57 cdent: oh, don't worry. It just means I may not be as dimwitted as usual 14:03:16 o/ 14:03:19 does that mean I should be holding on to my sides? 14:03:34 Hold on to whatever you like 14:04:16 FWIW, I'm still having 25 mins for you folks before rushing out 14:04:24 at least, till April 14:04:28 and then DST for me 14:04:47 o/ 14:05:28 Let's get started 14:05:31 #topic Specs & Reviews 14:05:52 traits ahoy 14:05:56 first up: Traits series, starting with: 14:06:06 #link 14:06:11 doh! 14:06:18 merged. 14:06:23 :) 14:06:24 #link https://review.openstack.org/#/c/376199/ 14:06:25 #link https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/resource-provider-traits 14:06:39 the patch series :) 14:07:10 alex_xu: this looks like it's getting very close 14:07:10 yup, ^^ it highest priority. 14:07:20 edleafe: thanks 14:07:21 alex_xu: Any issues we need to focus on or discuss? 14:07:25 alex_xu: I'm reviewing that series this morning. 14:07:42 1. I removed the cache 14:08:04 edleafe, alex_xu: I will have a couple patches up today for adding networking traits to os-traits. 14:08:05 2. is there anyway I can separate one microversion into two separated patches? 14:08:26 alex_xu: not if both patches add/change something in the API. 14:08:31 jaypipes: cool 14:08:45 alex_xu: but you can split out patches and isolate only the API-affecting stuff into a single patch. 14:08:46 jaypipes: ok 14:08:48 alex_xu: +1 to cache removal 14:09:34 ok, let me merge the last two API patches 14:09:48 jaypipes: those changes won't conflict with anything in alex_xu's series, will they? 14:11:27 Oh, well, guess Jay got pulled in another direction... :) 14:11:30 Moving on... 14:11:31 Ironic inventory is reeeeeally close: 14:11:33 #link https://review.openstack.org/#/c/441544/ 14:11:56 Let's all get our eyes on that 14:12:19 Nested RPs needs some attention: 14:12:20 #link https://review.openstack.org/#/c/415920/ 14:12:32 But those are waiting for the traits stuff 14:12:52 Finally: 14:12:53 Improve placement API error messages: 14:12:53 #link https://review.openstack.org/#/c/418393/ 14:13:11 Comments on any of these? 14:13:17 edleafe: you mean the os-traits ones? 14:13:27 jaypipes: yes 14:13:29 edleafe: sorry, no, they won't affect alex's patches. 14:13:44 jaypipes: cool, I didn't imagine they would 14:13:55 <_gryf> edleafe, there was some comments from stephen 14:14:05 <_gryf> which was addressed 14:14:14 _gryf: on which? 14:14:28 <_gryf> edleafe, imporve api error mesgs 14:15:16 I have in my queue to re-review with those changes 14:15:23 <_gryf> cool, hanks 14:15:27 <_gryf> *thanks 14:15:36 Anything else for Specs and Reviews? 14:15:56 #topic Bugs 14:16:05 No new bugs that I'm aware of 14:16:20 * edleafe hasn't searched that deeply, though 14:17:00 I looked late last week, and agree 14:17:13 Moving on... 14:17:17 #topic Open discussion 14:17:19 (I try to regularly look for things that need a placement or scheduler tag) 14:17:28 Functional testing: 14:17:29 https://etherpad.openstack.org/p/nova-placement-functional 14:17:40 We have some general approaches there 14:17:53 I have some first baby steps outlined at the bottom 14:18:15 I'd like to start working on this this week 14:18:25 Unless anyone has objections 14:18:44 not sure I get the problem 14:18:47 edleafe: ^ 14:19:13 edleafe: those baby steps lgtm 14:19:29 edleafe: you know that we have a PlacementFixture, right? 14:19:37 bauzas: we lack functional tests, and that prevents us from making changes with any confidence 14:19:37 * jroll reminds folks that he'll be writing specs for some of the flavor changes and ironic transition this week 14:19:58 bauzas: yes 14:20:14 jroll++ 14:20:15 bauzas: I'd like to use it more 14:20:41 edleafe: okay, you confused me with "jay's fixtures" then 14:21:21 and yeah, I agree with you, we're missing functional coverage 14:21:49 bauzas: the idea was to have a single base instead of several 14:22:59 So I'll probably start on this today, and have a first attempt soon. I'll ping people on -nova to review 14:23:22 Any other comments/concerns about functional testing changes? 14:24:09 whoops, forgot to 14:24:11 #link Functional testing: 14:24:12 https://etherpad.openstack.org/p/nova-placement-functional 14:24:29 * edleafe is having copy/paste issues this morning! 14:24:48 #undo 14:24:49 Removing item from minutes: #link https://etherpad.openstack.org/p/nova-placement-functional 14:24:53 #link https://etherpad.openstack.org/p/nova-placement-functional 14:25:17 edleafe: cool with me. ++ 14:25:38 OK, moving on. 14:25:39 Extracting placement from Nova 14:25:40 Dev ML discussion starting with: 14:25:41 #link http://lists.openstack.org/pipermail/openstack-dev/2017-March/113812.html 14:25:53 cdent: wanna run with this one? 14:26:29 a) the email pretty much says it: I extracted a working placement and found some few small issues, some of which we can deal with independently of any actual plan to do extraction 14:27:01 b) it wasn't my intent to such any timeline or solid plan, just to change guessing to less guessing, because all the guessing was causing some FUD 14:27:25 ++ 14:27:37 edleafe: on functional testing (sorry, late), please get with sfinucan, since he's also been working on that for libvirt. 14:28:00 c) that said, I continue to think that sooner is better than later: we have a self-contained thing here that is _easy_ to extract (unlike the rest of nova) and we can help both nova and placement by having disconnected review processes 14:28:03 EOS 14:28:21 jaypipes: will do 14:29:06 * bauzas bails out 14:29:17 damn kids 14:29:18 cdent: separating it sooner will mean less time for rust to build that will make them harder to separate later 14:29:23 cdent: I agree with the sentiment, as mentioned on the ML. I think getting the discussion started *now* about a separation in Queens is A Good Thing. dansmith has some concerns on whether to include claims in the list of things to do *before* split-out, but is willing to discuss further about an appropriate delimiting point. 14:29:28 rust == linkages 14:30:38 OK, how about we continue the discussion on the ML, and talk further at our next meeting? 14:30:44 +1 14:31:08 It's much too early in the thought process to be able to make any decisions, anyway 14:31:12 yes 14:31:19 <_gryf> btw, when is the spec freeze? 14:31:39 But I do think knowing the issues can help us clean up those bits earlier, so thanks cdent 14:31:51 * cdent bows 14:32:06 _gryf: https://releases.openstack.org/pike/schedule.html 14:32:11 <_gryf> thx 14:32:38 Anything else for Opens? 14:34:25 OK, thanks everyone! 14:34:27 #endmeeting