14:00:15 <edleafe> #startmeeting nova_scheduler
14:00:16 <openstack> Meeting started Mon May  7 14:00:15 2018 UTC and is due to finish in 60 minutes.  The chair is edleafe. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:19 <openstack> The meeting name has been set to 'nova_scheduler'
14:00:33 <efried> ō/
14:00:34 <takashin> o/
14:00:35 <tssurya> o/
14:00:44 <edleafe> #link Agenda https://wiki.openstack.org/wiki/Meetings/NovaScheduler#Agenda_for_next_meeting
14:00:47 <melwitt> o/
14:01:02 <tetsuro> o/
14:01:16 <sundar> o/
14:01:40 <edleafe> #topic Specs & Reviews
14:01:55 <edleafe> #link latest Placement update http://lists.openstack.org/pipermail/openstack-dev/2018-April/129941.html
14:02:14 <edleafe> That's a week old, as cdent had the nerve to take some time off
14:02:53 <efried> sheesh
14:02:58 <edleafe> Does anyone have any specs or reviews to discuss?
14:03:00 * bauzas waves hand
14:03:01 <jaypipes> o/
14:03:17 <bauzas> small week for me
14:03:51 <alex_xu> o/
14:03:51 <edleafe> bauzas: are you in SF?
14:03:57 <bauzas> nope at all
14:04:20 <bauzas> our Summit is for *customers* :p
14:04:29 <edleafe> ...and partners
14:04:33 <edleafe> most of my team is there
14:04:34 <bauzas> and I don't pay subscriptions :p
14:05:05 <bauzas> edleafe: opefully, I could finally pay my debt on you in Vancouver
14:05:13 <bauzas> hopefully* even
14:05:28 <bauzas> I have a memory that tells me I owe you a free beer
14:05:43 <edleafe> I don't remember that, but I'll certainly take a free beer!
14:06:35 <edleafe> OK, back to business. Nothing to discuss for Specs and Reviews?
14:07:06 <edleafe> #topic Bugs
14:07:10 <sundar> edleafe: what is the status if nested resource providers?
14:07:14 <edleafe> #link Placement bugs https://bugs.launchpad.net/nova/+bugs?field.tag=placement&orderby=-id
14:07:30 <edleafe> sundar: I'll those who are working on it answer
14:08:05 <efried> Series starting here: https://review.openstack.org/#/c/556450/
14:08:13 <efried> The series is complete, needs reviews.
14:08:47 <tetsuro> ready for review :)
14:09:01 <jaypipes> tetsuro: will review the series this morning.
14:09:18 <tetsuro> thanks in advance.
14:09:44 <efried> Granular series is two patches deep at this point, starts here: https://review.openstack.org/#/c/564351
14:10:16 <efried> There's a little contention as to whether we should do jaypipes' refactor (https://review.openstack.org/#/c/566166/) before or after ^
14:10:34 <jaypipes> efried: after.
14:10:42 <jaypipes> efried: no real contention.
14:10:44 <efried> cool, wfm.
14:11:07 <jaypipes> efried: gonna do tetsuro's patch series first, then yours. one of you will be rebasing anyway.
14:11:20 <bauzas> efried: planning to review https://review.openstack.org/#/c/556450/ hopefully tomorrow
14:11:24 <efried> argh.  Yeah, the loser has to figure out how to do nested + granular.
14:12:04 <efried> which really shouldn't be a huge deal code-wise - but the tests...
14:12:11 <bauzas> one ring to rule'em all
14:12:17 <jaypipes> tetsuro: your thoughts on https://review.openstack.org/#/c/566166/ would be good, too, since you're quite familiar with that area of the codebase.
14:12:31 <efried> jaypipes: I saw tetsuro left a review on that.
14:12:37 <jaypipes> ah, cool.
14:12:39 <efried> One step ahead of us, as usual :)
14:12:44 <jaypipes> :)
14:13:29 <efried> sundar: So hopefully we can have both of those things merged in the next week or so.
14:13:54 <bauzas> that's my intent to do such *before* the summit
14:14:13 <jaypipes> sundar: does Cyborg have its implementation of update_provider_tree() done yet?
14:14:45 <efried> What would be really neat is to get it done in time for me & edleafe to work it into the "functional" part of our presentation for Vancouver.
14:15:28 <bauzas> just one thing
14:15:30 <bauzas> I saw https://review.openstack.org/#/c/560444/2
14:15:45 <bauzas> which is the next step of https://review.openstack.org/#/c/564215/
14:16:34 <bauzas> are folks happy if I'm rebasing https://review.openstack.org/#/c/560444/ on top of https://review.openstack.org/#/c/564215/ ?
14:17:08 <efried> bauzas: I was just talking to lei-zh about https://review.openstack.org/560317 which needs to be rebased on top of that
14:17:10 <bauzas> AFAICS, https://review.openstack.org/#/c/560459/ doesn't have any bug report attached to it
14:17:27 <bauzas> mmmmm
14:17:41 <efried> bauzas: I've asked a couple of times and nobody has been able to tell me that there's a bug associated with that issue.
14:17:43 <bauzas> ok, so I'll rebase your change at the bottom of my series
14:17:53 <bauzas> and let it merge soon
14:17:58 <efried> Which seems weird to me, because my understanding was that that issue was the whole impetus for sharing providers in the first place.
14:18:18 <bauzas> efried: you had comments on testing, I'll try to fix those
14:19:03 <efried> bauzas: For the DISK_GB patch?  bhagyashris is working on it - pastebin'd a functional test in -nova a few minutes ago that's not quite working yet.
14:19:19 <bauzas> efried: no, about https://review.openstack.org/#/c/560444/2/nova/tests/unit/virt/libvirt/test_driver.py
14:19:42 <bauzas> efried: I'll address those comments and upload a new revision under my branch
14:20:09 <efried> bauzas: Okay, works for me, you may want to coordinate with bhagyashris.
14:20:19 <bauzas> hope it will gain more visibility than bhagyashris's patch that is neither matching a feature nor a bug
14:20:46 <bauzas> efried: do you know his IRC nick ?
14:20:53 <sundar> efried and all: Thanks!
14:20:59 <sundar> jaypipes: No, not yet
14:20:59 <bauzas> nvm, found
14:22:38 <efried> So bauzas are you also going to reshuffle lei-zh's patch in the mix?
14:22:40 <bauzas> aye, I need to leave
14:23:06 <bauzas> efried: I don't see https://review.openstack.org/#/c/560444/ being a dependency on more than just https://review.openstack.org/#/c/560459/3
14:23:48 <bauzas> efried: https://review.openstack.org/#/c/560317/ could be rebased on top of https://review.openstack.org/#/c/560444/ without needing a new revision
14:23:54 <efried> bauzas: hmu later and we can sort out the details.
14:24:02 <bauzas> yeah
14:24:23 <efried> https://review.openstack.org/#/c/560317/ needs to be reworked to ditch the get_traits compute driver method and use update_provider_tree instead.  Which means it should be on top of 444
14:24:23 <bauzas> anyway, the sooner we can merge https://review.openstack.org/#/c/560444/2, the better it will be
14:24:32 <efried> I'm on board with that.
14:24:39 <bauzas> agreed.
14:24:48 <bauzas> anyway, needs to taxilift my kids
14:24:50 <bauzas> \o
14:25:22 <edleafe> Ready to move on?
14:25:33 <alex_xu> efried: I guess 560417 just need to change the get_traits as a private method like `_get_traits`, and then calling it in update_provider_tree?
14:25:50 <efried> alex_xu: Yes.  Or even fold _get_traits into the util method.
14:26:30 <alex_xu> efried: ok, I feel fold that method a little mess for the update_provider_tree, but anyway let us review in the patch
14:26:48 <efried> yeah, I'm fine either way.  Just as long as we're not adding a new compute driver method that we don't need.
14:27:02 <efried> (ComputeDriver base class method, that is)
14:27:31 <alex_xu> yea, agree that shouldn't be a new driver method also
14:28:09 <efried> okay, let's move on.
14:28:28 <edleafe> #topic Open Discussion
14:28:56 <edleafe> I'd like to start by saying that after two years of running these meetings, I am stepping down
14:29:33 <edleafe> My work has focused me more on internal projects, and so I won't be able to continue to be as involved in upstream as before
14:30:17 <edleafe> Anything else for Open Discussion?
14:30:35 * efried is saddened by this news
14:30:52 <edleafe> OK, thanks everyone!
14:30:56 <edleafe> #endmeeting