14:01:15 <edmondsw> #startmeeting PowerVM Driver Meeting
14:01:16 <openstack> Meeting started Tue Aug 21 14:01:15 2018 UTC and is due to finish in 60 minutes.  The chair is edmondsw. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:19 <openstack> The meeting name has been set to 'powervm_driver_meeting'
14:01:31 <edmondsw> #link agenda: https://etherpad.openstack.org/p/powervm_driver_meeting_agenda
14:02:18 <edmondsw> #topic In-Tree Driver
14:03:16 <esberglu> o/
14:03:28 <edmondsw> I've been thinking about what we should attempt to do in-tree for Stein
14:04:17 <edmondsw> we've talked about maybe doing some of the device passthrough work both IT and OOT in Stein, so that's one thing
14:04:35 <edmondsw> the other obvious thing is the work that we didn't contain in Rocky for resize, migration, evacuation
14:04:42 <edmondsw> all the things that need multi-node CI
14:04:54 <edmondsw> anyone want to throw out other ideas?
14:05:03 <edmondsw> or comment on those?
14:06:01 <edmondsw> don't everyone talk at once... ;)
14:06:21 <edmondsw> ok, we don't have to hold the meeting on that... but think about it, and let me know
14:06:35 <efried> .
14:06:44 <edmondsw> anything else to discuss for IT today?
14:06:46 <efried> Are there any prerequisites in-tree
14:06:48 <efried> for device passthrough
14:06:51 <efried> to be implemented in-tree?
14:07:06 <efried> I'm not coming up with any offhand.
14:07:12 <edmondsw> not that I know of
14:07:20 <efried> And in fact it may be easier to implement passthrough before we've implemented things like live migration and remote restart.
14:07:48 <edmondsw> at this point, the OOT driver is still the "full-feature" driver
14:07:54 <edmondsw> so anything we do IT we need to also do OOT
14:08:08 <edmondsw> so I don't know that it will make anything easier to do IT first
14:08:20 <edmondsw> when OOT would have to follow immediately in the same release
14:08:50 <edmondsw> unless it's a question of getting nova core eyes on it, which does have significant benefits
14:09:14 <edmondsw> (not having to reimplement OOT to match IT if we have to make a lot of changes to satisfy them IT)
14:10:05 <edmondsw> but let's say we do IT first... we need to do that with an eye on how we'll handle things like live migration and RR when we implement OOT
14:10:21 <edmondsw> and probably have those patches up in parallel
14:10:29 <edmondsw> efried thoughts?
14:10:37 <edmondsw> feel like I'm typing to myself :)
14:10:57 <efried> I was thinking about doing them in parallel, yes.
14:11:15 <efried> Especially if the in-tree work that kosamara is driving gets any traction.
14:11:15 <edmondsw> makes sense to me
14:11:29 <efried> I want to be the first driver to do it.
14:11:35 <efried> or first-with-libvirt.
14:11:44 <edmondsw> +1 to that
14:12:51 <edmondsw> alright, moving on
14:13:01 <edmondsw> #topic Out-of-Tree Driver
14:13:03 <efried> otherwise, I don't have any particular agenda wrt working towards parity
14:13:36 <edmondsw> https://review.openstack.org/#/c/593529/ needs a 2nd +2 and +A
14:14:47 <edmondsw> https://review.openstack.org/#/c/592057/ already has 2 +2's but I'm happy leaving that open until gautpras gets back from vacation tomorrow
14:14:55 <edmondsw> I'd like to have his vote on that before merging
14:15:47 <edmondsw> mdrabe I've made some progress on stacking a multinode environment for you (for MSP testing)... and then screwed it up somehow... so still working on that
14:16:19 <edmondsw> that covers the efforts I'm currently aware of IT... anything else?
14:16:33 <edmondsw> sorry... meant OOT
14:17:18 <edmondsw> I am keeping an eye on https://review.openstack.org/#/c/592520/ and will probably propose similar changes to our .gitignore files just to clean them up
14:18:12 <edmondsw> #topic Device Passthrough
14:18:14 <edmondsw> efried ^
14:18:56 <efried> Other than some more design discussion on kosamara's spec, nothing new to report since last week.
14:20:51 <edmondsw> alright, tx
14:20:57 <edmondsw> #topic PowerVM CI
14:21:26 <edmondsw> we had some CI issues tracing back to the zuul merger nodes, but I believe that's resolved and things are working now
14:21:43 <edmondsw> mujahidali over to you
14:22:46 <mujahidali> esberglu: talked about getting help from tonyb for disk-image-builder.
14:23:02 <edmondsw> as part of the nodepool upgrade, yes
14:23:11 <edmondsw> which will be a prereq for multinode CI at this point
14:25:04 <edmondsw> mujahidali are you working on that in the staging environment?
14:25:26 <edmondsw> or should we make the change that esberglu has suggested for vSCSI in 6596 and try that in staging?
14:25:45 <edmondsw> I'd also like to try 6684 in staging
14:25:54 <mujahidali> yes
14:26:00 <mujahidali> I was looking into this https://github.com/openstack-infra/nodepool/blob/0.3.0/doc/source/configuration.rst
14:26:29 <mujahidali> pointed by esberglu
14:28:26 <edmondsw> mujahidali ok if you are making progress on multinode, we can leave the staging environment tied up with that
14:28:33 <edmondsw> it's more important than the other 2 things I mentioned
14:29:16 <mujahidali> I wanted to try that as well in parallel but got stuck with py2-py3 work.
14:29:35 <edmondsw> sure
14:29:45 <edmondsw> anything else to discuss for CI today? esberglu mujahidali
14:30:18 <esberglu> Nothing from me
14:30:25 <mujahidali> we are waiting for the devstack stable/rocky ??
14:30:32 <edmondsw> ah, yes
14:30:39 <mujahidali> okay
14:30:42 <edmondsw> I've not gotten a reply from the swift guys
14:30:53 <edmondsw> and devstack is waiting for them
14:31:27 <edmondsw> but I think we can afford to wait a couple days if need be
14:31:34 <mujahidali> sure
14:31:55 <esberglu> edmondsw: I had one comment on 6684
14:32:22 <edmondsw> tx
14:33:06 <edmondsw> #topic Open Discussion
14:33:23 <edmondsw> opening up the floor... whatchy'all got?
14:33:27 <mujahidali> I will be OOO on 23rd and 24th.
14:33:37 <edmondsw> mujahidali ack
14:33:54 <edmondsw> I will be out next Mon-Thurs
14:34:10 <edmondsw> so no meeting next week unless someone else wants to volunteer to run it
14:35:48 <edmondsw> alright, thanks everyone
14:35:52 <edmondsw> #endmeeting