09:00:02 <anteaya> #startmeeting nova-net-to-neutron-migration
09:00:02 <openstack> Meeting started Tue Mar  3 09:00:02 2015 UTC and is due to finish in 60 minutes.  The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:05 <openstack> The meeting name has been set to 'nova_net_to_neutron_migration'
09:00:08 <obondarev> hi
09:00:09 <anteaya> hello
09:00:15 <anteaya> how are you oleg?
09:00:25 <anteaya> gus are you around?
09:00:31 <anteaya> mikal ping
09:00:34 <obondarev> I'm great! Thank you :)
09:00:44 <anteaya> I don't see belmoria in channel
09:00:51 <anteaya> obondarev: glad to hear it
09:00:57 <anteaya> might just be you and I
09:01:14 <obondarev> anteaya: I see
09:01:21 <anteaya> shall we get started?
09:01:32 <obondarev> anteaya: is there some conference or smth?
09:01:35 <anteaya> #topic the state of the Neutron spec (obondarev)
09:01:40 <obondarev> anteaya: sure we can
09:01:54 <anteaya> obondarev: no I think we are just close to feature freeze for kilo
09:02:03 <anteaya> so folks are otherwise occupied
09:02:08 <anteaya> it will pick up again
09:02:16 <mikal> Hi
09:02:16 <anteaya> so any change on spec status this week?
09:02:21 <anteaya> mikal: you are here
09:02:22 <obondarev> ok, I see
09:02:23 <anteaya> yay
09:02:27 <obondarev> no updates on the specs side again
09:02:31 <anteaya> moving on
09:02:32 <obondarev> so probably we should move to implementation topic
09:02:42 <anteaya> #topic the state of implementation (obondarev)
09:02:47 <anteaya> go ahead
09:02:55 <obondarev> I tested jlibosva's db migration patch a little bit on devstack
09:03:03 <obondarev> it seems working fine for flat (and flatdhcp, no floating ips) for my really simple data
09:03:04 <anteaya> wonderful
09:03:11 <anteaya> terrific
09:03:14 <obondarev> there is a couple of nits which I'll comment on review
09:03:30 <anteaya> would you be able to make a comment on the patch stating under what conditions it worked for you?
09:03:37 <anteaya> and yes, thank you
09:03:44 <obondarev> anteaya: sure
09:03:51 <obondarev> will comment that
09:04:08 <obondarev> for nova-net proxy I've uploaded new patch set last week
09:04:37 <obondarev> so in its current state it should allow to handle allocate/deallocate_for_instance in proxy mode
09:04:40 <anteaya> I see that thank you
09:04:45 <anteaya> so far no reviews
09:04:55 <anteaya> were you able to read the logs from the last nova meeting?
09:05:14 <obondarev> I think I red
09:05:34 <anteaya> so you caught that dan smith is still not sold that this is the best direction?
09:05:36 <obondarev> *read
09:05:44 <obondarev> yeah
09:05:48 <anteaya> okay
09:05:59 <obondarev> I know he thinks that the best is to not do it at all
09:06:12 <anteaya> not to do a proxy at all?
09:06:19 <obondarev> yes
09:06:33 <anteaya> what do you think he would support as a direction?
09:06:49 <obondarev> I don't know
09:06:57 <anteaya> mikal: any insight here?
09:07:31 <anteaya> what do you think dan and nova would support as a direction, since we don't seem to be getting support for a proxy
09:07:53 <mikal> So the short answer is I am not sure
09:07:59 <mikal> I need ti sit down and tlak with Dan about it more
09:08:04 <mikal> But I haven't found the time for that
09:08:08 <obondarev> I feel Dan's main concern is that with the proposed approach we'll not be able to handle complex cases
09:08:13 <mikal> I accept this is a thing I need to find time for though
09:08:44 <obondarev> currently I'm debugging proxy on devstack, found several issues there
09:08:53 <obondarev> going to fix them and upload a working version
09:08:58 <obondarev> today or tomorrow
09:09:22 <anteaya> okay so if we don't have support for the current direction, then perhaps we can work on finding support for some direction to be determined
09:09:24 <anteaya> yes?
09:09:48 <obondarev> I guess so
09:10:00 <mikal> anteaya: remind me of the review number for the nova proxy?
09:10:04 <obondarev> otherwise I don't know what's the way forward
09:10:07 <anteaya> sorry
09:10:21 <anteaya> #link https://review.openstack.org/#/c/150490/
09:10:22 <anteaya> proxy
09:10:32 <mikal> Ta
09:10:36 <anteaya> #link https://review.openstack.org/#/c/148260/
09:10:40 <anteaya> db migration
09:10:43 <anteaya> thanks for asking
09:10:49 <obondarev> I'll also add a comment with an instruction on how it can be tested together with db migration patch
09:10:55 <obondarev> with what to do and what to expect
09:11:13 <obondarev> hope this will be enough to prove that the approach does its job
09:11:26 <mikal> So, re-reading that Dan'
09:11:30 <mikal> s concern is quite specific
09:11:38 <mikal> He doesn't want to fake out the sqlalchemy layer
09:11:55 <mikal> "I feel like doing this at the nova/network/base_api.py layer is the proper place, probably subclassing the nova-net one and re-using a lot of what is there. I guess that I don't believe that just replacing the persistence layer is going to be enough to cover more than just the most simple of cases."
09:12:04 <obondarev> yeah
09:12:29 <obondarev> that's why I prefer that all the discussion should happen on the review
09:12:36 <mikal> But yeah, let me chat with Dan and let you know what I can find out from a high bandwidth conversation next week
09:12:48 <obondarev> we discussed that with Dan on irc
09:12:56 <mikal> Oh, ok
09:12:59 <mikal> Where did that end up?
09:13:16 <obondarev> so he said he can't come up with a better approach
09:13:28 <obondarev> rather that don't do it
09:14:16 <obondarev> in my comments I tried to explain why proxying at api layer is not simpler and clearer
09:14:20 <mikal> But what about the base_api comment
09:14:24 <mikal> Is that his proposed better approach?
09:15:14 <obondarev> that was before our conversation
09:15:36 <obondarev> probably I convinced him
09:15:43 <obondarev> not sure however
09:15:48 <mikal> Ok
09:15:50 <anteaya> obondarev: you didn't convince him
09:15:57 <mikal> Like I said, let me chat to him and see what I can learn
09:16:02 <anteaya> since his feeling at the nova meeting remained the same
09:16:06 <obondarev> but he seem not advocating api approach any more
09:17:08 <obondarev> anteaya: from what I saw his only suggestion was not do it
09:17:16 <anteaya> right
09:17:27 <anteaya> which means we need to find out what to do
09:17:28 <obondarev> anteaya: he didn't mention any other approach, did he?
09:17:40 <mikal> So, I think we should table this until I can chat with Dan
09:17:47 <mikal> We seem to now be trying to guess what he thinks
09:17:49 <anteaya> I didn't ask, since i knew you wanted to keep working on the proxy
09:17:50 <mikal> Which seems unproductive
09:17:58 <anteaya> mikal: good point
09:18:00 <obondarev> anteaya: tha't why I thought I probably convinced him that api approach isn't better
09:18:02 <anteaya> moving on
09:18:08 <anteaya> you didn't
09:18:29 <anteaya> #topic documentation
09:18:36 <anteaya> so nothing new here
09:19:08 <anteaya> we have a patch up to docs but it is on hold while they go from xml to rst
09:19:21 <anteaya> #topic testing
09:19:34 <anteaya> so we have the db migration if anyone wants to test it
09:19:39 <anteaya> anything more in testing?
09:20:43 <anteaya> okay moving on
09:20:51 <anteaya> #topic open discussion
09:21:04 <anteaya> anyone have anything here?
09:21:18 <anteaya> okay let's call it then
09:21:29 <anteaya> obondarev mikal thanks for your attendance, I'm grateful
09:21:40 <anteaya> mikal: have fun tomorrow on your visit to the salt mines
09:21:43 <obondarev> anteaya: thank you for organizing
09:21:46 <anteaya> sure
09:21:56 <anteaya> I look forward to seeing where we are next week
09:22:06 * gus shuffles in quietly and hopes no-one notices his tardiness...
09:22:06 <anteaya> hopefully we can have some directional guidance
09:22:11 <anteaya> gus: hi
09:22:14 <anteaya> we are just finished
09:22:17 <mikal> anteaya: thanks!
09:22:27 <anteaya> gus: do you have anything you wish to add before we close out?
09:22:32 <mikal> gus: well timed!
09:22:56 <gus> heh, no I have nothing to add this week either.
09:23:02 <anteaya> okay thanks
09:23:10 <anteaya> have a good rest of time of day to you
09:23:13 <anteaya> see you next week
09:23:18 <anteaya> #endmeeting