14:00:31 <edmondsw> #startmeeting PowerVM Driver Meeting 14:00:32 <openstack> Meeting started Tue Jun 5 14:00:31 2018 UTC and is due to finish in 60 minutes. The chair is edmondsw. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:36 <openstack> The meeting name has been set to 'powervm_driver_meeting' 14:00:40 <esberglu> o/ 14:00:44 <edmondsw> #link agenda: https://etherpad.openstack.org/p/powervm_driver_meeting_agenda 14:01:19 <edmondsw> ping efried gman-tx mdrabe mujahidali chhagarw 14:01:29 <edmondsw> #topic In-Tree Driver 14:01:37 <edmondsw> #link https://etherpad.openstack.org/p/powervm-in-tree-todos 14:01:55 <esberglu> Nothing from me 14:02:14 <edmondsw> esberglu really? you don't want to brag about getting vSCSI merged? 14:02:40 <esberglu> edmondsw: Was that not in last meeting? 14:02:59 <esberglu> Yeah vSCSI merged, the snapshot cleanup merged, we're off the runway 14:03:20 <esberglu> I'm still compiling backports to OOT, should have a good list together this week 14:03:21 <edmondsw> it was late last week, so since the last meeting 14:03:45 <edmondsw> tx 14:04:03 <edmondsw> #topic Out-of-Tree Driver 14:04:12 <edmondsw> #link https://etherpad.openstack.org/p/powervm-oot-todos 14:04:13 <efried> ō/ 14:05:21 <edmondsw> chhagarw has been struggling to get a multinode devstack env stacked to test iSCSI migration cases 14:05:35 <edmondsw> esberglu and I have been helping (mostly esberglu) 14:06:02 <esberglu> edmondsw: Control node is up, and I think the compute node _should_ stack now 14:06:14 <edmondsw> +1 14:06:25 <esberglu> edmondsw: As a side note we had to bump the timeout waiting for the compute service to come up from 60 to 720 14:06:41 <edmondsw> any idea why it needs to be that long in her env? 14:06:44 <esberglu> Might be something environmental I've never seen it take that long 14:07:00 <esberglu> But I don't have any good ideas for why 14:09:05 <edmondsw> I wonder if it goes back to https://review.openstack.org/#/c/471773/ 14:09:24 <edmondsw> at any rate, glad it stacked 14:09:52 <edmondsw> efried will you be able to update the lower-constraints.txt for https://review.openstack.org/#/c/555964/ ? 14:10:47 <efried> I could. Got a bunch going on at the moment. 14:11:16 <efried> Not 100% sure there's an obviously-correct way to do that, though. 14:11:16 <edmondsw> no rush 14:11:36 <edmondsw> yeah, I think what we put there can be somewhat flexible 14:11:46 <efried> it seems weird/wrong to have to duplicate nova's entire l-c. 14:11:46 <edmondsw> but you agree it makes sense to match nova as much as possible, right? 14:11:53 <efried> match where common, sure. 14:12:10 <edmondsw> I agree that does seem weird. If we can figure out how to pull their file into ours dynamically... 14:12:18 <efried> yeah, was just going to say that. 14:12:22 <efried> That's totally possible. 14:12:37 <edmondsw> what we need to add should be a fairly small list 14:13:18 <edmondsw> tjakobs you think https://review.openstack.org/#/c/569430/ will be unblocked soon? 14:13:25 <efried> on top of what nova needs? Should be nil, really. 14:13:36 <edmondsw> efried could be 14:14:19 <tjakobs> edmondsw: the REST API code should be done soon (just need a +2) 14:14:48 <edmondsw> we're gonna need a new pypowervm release cut so we can pull that into our requirements 14:15:03 <edmondsw> for that and for https://review.openstack.org/#/c/568374/ 14:15:17 <edmondsw> tjakobs you working on anything else for nova-powervm or just those 2? 14:15:56 <tjakobs> edmondsw: not sure we need pypowervm for the rbd resize, but we do for the LOOP 14:16:10 <tjakobs> I'm just working on those (plus kinda the refactor eventually) 14:16:49 <edmondsw> yeah, when we can 14:16:59 <edmondsw> ok, anything else to discuss for OOT anyone? 14:17:23 <edmondsw> #topic Device Passthrough 14:17:29 <edmondsw> efried ? 14:17:52 <efried> Big discussion on upgrade path for virt drivers moving to modeling via nrp. 14:18:03 <efried> hangout in 45m to discuss 14:18:20 <efried> Jay tried using it as an excuse to stop reviewing the nrp-in-alloc-cands series. I shut that down. 14:18:55 <efried> That's still out there. I've +2ed the bottom couple, and some in the middle; waiting for Jay to show some action and I'll hit the rest. 14:19:06 <edmondsw> should I join the hangout, or you got it? 14:19:21 <efried> If you're interested, but no need. 14:19:28 <edmondsw> k, I'll leave it to you 14:20:10 <openstackgerrit> Eric Fried proposed openstack/nova-powervm master: add lower-constraints job https://review.openstack.org/555964 14:20:12 <efried> Let's see if that works ^ 14:20:25 <efried> removed our l-c entirely, just using nova's. 14:20:30 <efried> cause that actually makes sense. 14:20:41 <edmondsw> +1 14:20:43 <efried> and lets us move with theirs. 14:22:05 <openstackgerrit> Eric Fried proposed openstack/nova-powervm master: add lower-constraints job https://review.openstack.org/555964 14:22:23 <efried> rebase ^ 14:22:49 <edmondsw> efried anything else for passthrough 14:23:02 <efried> so the way this upgrade business affects us is: we should wait until nrp-in-alloc-cands merges before we start working on the code to model our providers. 14:23:06 <efried> ...as trees 14:23:11 <efried> ...which we were going to do anyway. 14:23:13 <efried> but 14:23:19 <efried> if everything goes to hell 14:23:43 <efried> the upgrade thing gives us a way to model stuff non-nested for now, and then allow us to reshape later. 14:24:13 <edmondsw> interesting 14:24:28 <efried> so basically if we start feeling like nrp-in-alloc-cands is at too much risk, we can fake it with resource classes, or sharing providers, or whatever. 14:24:34 <efried> and then "fix" it later. 14:24:48 <edmondsw> cool 14:24:50 <efried> that is all 14:24:56 <edmondsw> #topic PowerVM CI 14:25:07 <edmondsw> #link https://etherpad.openstack.org/p/powervm_ci_todos 14:25:13 <edmondsw> mujahidali esbergu ^ 14:26:09 <edmondsw> PowerVM CI looks healthy at http://ci-watch.tintri.com/project?project=nova 14:26:34 <esberglu> mujahidali: Have you been monitoring the CI failures? Anything new? 14:27:02 <mujahidali> yeah 14:27:18 <esberglu> mujahidali: Great. Have you started looking into reinstalling those extra systems we received at all? 14:27:31 <esberglu> mujahidali: I haven't had a chance to look at your latest vSCSI runs, will do today 14:27:55 <mujahidali> no, I still need to get the system details from Drew. 14:28:02 <edmondsw> esberglu as for the extra systems, they weren't reachable but that's fixed now 14:28:03 <mujahidali> Stable/queens passed tempest but for stable/pike while running tempest it's giving import error 14:28:13 <edmondsw> I have the systems in my name now 14:28:26 <mujahidali> for the vscsi stable/branches 14:28:36 <edmondsw> I started trying to greenfield deploy NovaLink and hit some issues, had to put it aside 14:28:47 <esberglu> I'll be pushing up new patchsets for multinode CI either today or tomorrow. It's getting really close, I will detail what we have left in the commit 14:29:14 <edmondsw> mujahidali did you make sure you were using devstack's pike branch for testing pike? 14:29:49 <edmondsw> I think your issues might have been due to using devstack master 14:29:57 <edmondsw> esberglu that make sense? ^ 14:29:59 <mujahidali> I exported ZUUL_BRANCH=stable/pike 14:30:13 <edmondsw> the import errors were things that didn't exist in pike 14:30:27 <esberglu> edmondsw: Maybe, I only took a glance at it 14:32:25 <mujahidali> one more thing, I need to explicitly set the ETCD_VERSION=v3.2.0-rc.1 14:32:30 <edmondsw> mujahidali did you create a commit with your changes so far and post it up to gerrit like I suggested, so we can look? 14:32:55 <mujahidali> although in the local.conf file it's v3.2.0-rc.1 only 14:33:38 <edmondsw> sounds like you're doing something wrong 14:33:53 <esberglu> mujahidali: ETCD_VERSION should already be set to that for stable/pike runs 14:33:54 <edmondsw> get that commit created, and then let's talk through your steps in using it 14:34:01 <esberglu> https://github.com/powervm/powervm-ci/blob/master/devstack/stable/pike/outoftree/local.conf#L62-L65 14:34:03 <mujahidali> sure 14:34:28 <esberglu> So you're correct in having it set to that 14:35:05 <edmondsw> anything else for CI? 14:35:26 <esberglu> Not from me 14:35:53 <mujahidali> esberglu: you are correct, it should set v3.2.0-rc.1, but when I run without export it's picking 3.1.10 and when I look in local.conf file it's v3.2.0-rc.1 14:36:57 <edmondsw> alright let's debug that after the meeting once you have the commit up for review 14:37:04 <edmondsw> #topic Open Discussion 14:37:14 <edmondsw> anyone have anything? 14:37:46 <esberglu> I'm out thursday and friday 14:37:51 <edmondsw> ack 14:37:58 <efried> nova spec review day today 14:38:21 <mujahidali> I am out for next week 14:38:23 <edmondsw> Rocky-2 milestone is Thursday 14:38:40 <edmondsw> so I'll be proposing tags then 14:39:35 <edmondsw> alright, sounds like that's it for today. Thanks everyone 14:39:37 <edmondsw> #endmeeting