13:01:19 <esberglu> #startmeeting powervm_driver_meeting
13:01:20 <edmondsw_> esberglu meeting?
13:01:21 <openstack> Meeting started Tue Oct 31 13:01:19 2017 UTC and is due to finish in 60 minutes.  The chair is esberglu. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:01:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:01:25 <openstack> The meeting name has been set to 'powervm_driver_meeting'
13:01:27 <edmondsw_> :) o/
13:01:27 <efried> \o
13:02:05 <esberglu> #topic In-Tree Driver
13:02:07 <mdrabe> o/
13:02:42 <esberglu> Config Drive has a couple +1
13:03:04 <esberglu> Ready for core reviews? Or do we want thorst to take a look first?
13:03:18 <edmondsw_> I thought he'd looked already... checking
13:04:09 <edmondsw_> yeah, I guess his last review was quite a while ago
13:04:20 <edmondsw_> I think he'd be fine with us going after core reviews without him, though
13:04:27 <edmondsw_> I'll ping him so he's aware anyway
13:05:41 <edmondsw> thorst_ ^
13:05:57 <thorst_> edmondsw: I'll take a peak
13:06:12 <esberglu> OVS is still underway not much to report there
13:07:56 <edmondsw> esberglu think we'll be able to make some good progress there this week?
13:08:43 <esberglu> edmondsw: Yeah I think so
13:08:57 <edmondsw> cool
13:09:46 <edmondsw> after OVS we have SEA and vSCSI
13:10:02 <edmondsw> when do we have to have those in before they start locking things down?
13:10:03 <esberglu> Yep
13:10:22 <esberglu> Feature freeze is 1/25
13:10:44 <esberglu> But well before that to leave time for reviews
13:11:03 <edmondsw> yep... I was worried it would be earlier, that sounds fine
13:11:48 <edmondsw> that's probably all for IT then
13:12:20 <esberglu> edmondsw: I think we should target getting these done by the end of the year to leave a little extra space if needed
13:12:29 <edmondsw> esberglu agreed
13:12:57 <esberglu> #topic Out-of-Tree Driver
13:13:41 <esberglu> I still need to fix up UT on the nova-powervm side of the HostCPUStats stuff
13:13:51 <esberglu> Probably gonna do that this pm
13:14:44 <esberglu> Looks like I still have a little time before 1.1.9 though
13:15:44 <edmondsw> I think we're very close to getting iSCSI merged, which is the other big thing I wanted in for 1.1.9
13:16:05 <edmondsw> hopefully this week
13:17:19 <edmondsw> I don't have anything else OOT
13:17:28 <efried> I did a bug scrub yesterday
13:17:34 <edmondsw> efried tx for that
13:18:08 <efried> edmondsw and I almost came to blows over a poorly-worded message of mine.
13:18:18 <edmondsw> lol... s' ok
13:18:41 <edmondsw> I don't want to come to blows with you of all people...
13:18:49 <esberglu> lol
13:18:51 <efried> I don't hit.
13:19:03 <edmondsw> yeah, but there are other ways to bring the pain
13:19:11 <efried> Tap, snap, or nap, baby.
13:19:19 <efried> Moving on...
13:19:19 <esberglu> #topic Device Passthrough
13:19:44 <efried> Word on the street is we're no longer needing to hack PCI passthrough based on Pike.
13:20:06 <edmondsw> oh?
13:20:08 <efried> So I killed the respective patches and blueprints.
13:20:32 <efried> Yeah, decided it was gonna be too ambitious to try to do it in that timeframe, so we're waiting until the whatever-release-is-based-on-queens release.
13:20:46 <efried> Wherein we ought to be able to do things properly with generic resource providers.
13:21:04 <efried> I may even be able to have a prototype of the community side in a month or two.
13:21:06 <edmondsw> assuming we can use the nova scheduler...
13:21:35 <efried> mdrabe Were you looking into that side of things?
13:22:01 <edmondsw> it's being looked at, but I don't think anything definitive has been decided
13:22:03 <mdrabe> Nope
13:22:20 <edmondsw> at least, I saw a sizing
13:22:51 <edmondsw> we might still need those patches for queens, is my point. I hope not, but we might
13:23:27 <edmondsw> I'll see what I can figure out
13:23:54 <efried> Note that we needn't use the scheduler to make use of placement.
13:24:23 <efried> Anyway...
13:24:49 <efried> On the nova side, I've been developing the code for the granular resource syntax.  Got a couple series up for that.
13:25:11 <efried> Also been working with alex_xu on the placement side of GET /allocation_candidates
13:26:15 <efried> If you guys want to add reviews for the granular stuff, the flavor side is the series starting at https://review.openstack.org/#/c/515151/ and the placement side is the series starting at https://review.openstack.org/#/c/514091/
13:27:18 <edmondsw> mdrabe if you get a chance... ^
13:27:44 <efried> I'm also getting involved with how ironic is using traits.  That sorta came about organically as a result of me doing some spec reviews.  Not really relevant to powervm, except that it's interesting to see how they're modeling stuff and using placement.
13:28:03 <efried> Eventually I'll be using their stuff as an example of how to rework our driver accordingly.
13:28:24 <efried> For resources in general, that is.  They don't play device passthrough obviously.
13:28:25 <edmondsw> I think that's good. I would like us to have a better understanding of ironic
13:28:34 <edmondsw> I think that will be valuable going forward
13:28:45 <efried> thorst_ said he was looking at ironic too, though I don't know why or in what capacity.
13:29:04 <edmondsw> we can talk offline
13:29:28 <efried> That's it for me.
13:31:08 <esberglu> #topic PowerVM CI
13:31:28 <esberglu> hsien delivered a change that he thinks should fix that map LU issue
13:32:02 <edmondsw> that would be great
13:32:04 <esberglu> I deployed it on CI yesterday, gonna look through failures later and see if it helped and if anything else is affected
13:32:17 <edmondsw> I've been seeing a fair number of timeouts
13:32:38 <esberglu> edmondsw: Yep. I'm setting aside an afternoon this week to look at those
13:32:56 <edmondsw> really want to get this merged, but can't until CI passes: https://review.openstack.org/#/c/514858/
13:33:20 <edmondsw> esberglu if you could keep an eye on / keep trying that one
13:33:58 <esberglu> edmondsw: ack
13:34:57 <esberglu> That's it for me
13:35:04 <esberglu> #topic Open Discussion
13:35:23 <edmondsw> summit is coming up next week
13:35:48 <edmondsw> so if there's anyone you want me to catch while we're there, or something you want me to bring up or keep an ear out for...
13:36:04 <edmondsw> and I don't know if I'll make this meeting next week
13:36:14 <esberglu> Who's all going to that?
13:36:25 <efried> jaypipes isn't going, which I think will be a good opportunity for him & me to get some coding done.
13:36:26 <edmondsw> me, Joe, Amartey, Gerald
13:36:39 <esberglu> cool
13:39:38 <edmondsw> guess that's it
13:44:15 <edmondsw> got a +1 on the config drive patch from thorst_
13:45:14 <edmondsw> esberglu you want us to start pinging cores, or wait for you to address nits?
13:49:42 <esberglu> edmondsw: I'll hit the nits first
13:50:59 <edmondsw> esberglu k
13:59:45 <efried> #endmeeting
13:59:57 <efried> Guess esberglu has to do that.
14:11:46 <efried> esberglu Can you #endmeeting?
14:11:53 <esberglu> Oh sorry
14:11:56 <esberglu> #endmeeting