16:03:40 <rkukura> #startmeeting networking_ml2
16:03:42 <openstack> Meeting started Wed Nov  9 16:03:40 2016 UTC and is due to finish in 60 minutes.  The chair is rkukura. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:03:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:03:45 <openstack> The meeting name has been set to 'networking_ml2'
16:03:55 <rkukura> #topic Agenda
16:04:08 <rkukura> #link https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_November_9.2C_2016
16:04:28 <rkukura> Anything to add/remove?
16:04:52 <rkukura> #topic Announcements
16:05:04 <rkukura> I have no announcements - does anyone?
16:05:43 <rkukura> #topic Ocata Summit Retrospective
16:05:46 <Sukhdev> Nova's O-1 deadline is next week
16:05:59 <rkukura> thanks Sukhdev!
16:06:58 <rkukura> I was not at the summit, but I think Sukhdev, andreas_s, and maybe others were
16:07:11 <rkukura> who would liike to report?
16:08:22 <andreas_s> so what I learned, is that there are plans to deprecated l2pop
16:08:56 <rkukura> andreas_s: And replace it?
16:09:04 <rkukura> And what is the motivation?
16:09:43 <Sukhdev> So, I was in the session when this discussion took place
16:10:10 <Sukhdev> the idea is that the information provided by l2poop is not available in port object
16:10:35 <Sukhdev> the idea is use the port object to get that information instead of using l2poop
16:10:52 <Sukhdev> did I articulate it correctly, andreas_s ?
16:11:35 <Sukhdev> opps - I meant to say the information provided by l2pop IS available in port object
16:11:39 <andreas_s> so I got that they plan to make the relevant information available in the port object
16:11:46 <andreas_s> Sukhdev, yes
16:12:09 <rkukura> So is this another step of moving concepts from ML2 (like segmentation) into the Neutron core?
16:12:30 <andreas_s> kevinbenton is driving the l2pop deprecation effort - so If you need deeper insights, he is the right address ;)
16:13:15 <rkukura> There must have been some discussion of why it needs to move from ML2 to core. Is this so monolithic plugins can use it?
16:14:14 <andreas_s> let me try to find the summary mail from the summit..
16:14:50 <rkukura> andreas_s: thanks
16:16:34 <andreas_s> not mentioned in the session summary
16:16:56 <Sukhdev> andreas_s : there was an etherpad
16:17:50 <andreas_s> #link https://etherpad.openstack.org/p/ocata-neutron-agents
16:18:11 <andreas_s> the etherpad contains some of the information.
16:18:29 <andreas_s> just realized that this was part of the neutron agent session - there's not yet a summary for that one
16:19:24 <rkukura> ok
16:19:37 <rkukura> any other discussions that are relevant to ML2?
16:20:13 <andreas_s> the other thing was the reordering of the nova neutron communication during live migration
16:20:39 <andreas_s> nova convinced us to rather provide a genric portbinding api, rather than a live migration specific api
16:21:12 <rkukura> andreas_s: Is there any sort of proposal of what that API would look like?
16:21:12 <andreas_s> the argument was, that neutron should not be live migration aware
16:21:30 <andreas_s> yeah I updated the spec - this should come close to what was discussed at the summit
16:21:39 <andreas_s> #link https://review.openstack.org/309416
16:22:00 <andreas_s> but it still needs some review if this matches the expectations
16:22:50 <rkukura> andreas_s: thanks!
16:23:15 <andreas_s> in general - this new approarch raises a lot of questions, like how to deal with router bindings?
16:23:26 <andreas_s> I dumped them all into the spec
16:23:26 <rkukura> was there discussion around extensibility vs. versioning for this binding API?
16:23:40 <andreas_s> rkukura, no
16:24:03 <andreas_s> at least I can't remember
16:25:11 <rkukura> Sounds like this spec review is the right place for this discussion
16:25:17 <andreas_s> Other than that, the ovo work should go on
16:25:37 <rkukura> Were there any specific regarding OVO impact on ML2 driver APIs?
16:25:42 <andreas_s> and contract db migrations are not allowed anymore
16:25:50 <rkukura> Or on extension drivers?
16:26:17 <andreas_s> rkukura, no - I just saw some patches flying around, transforming some ML2 objects into ovos
16:26:18 <rkukura> is that ban on db contractions temporary?
16:27:02 <andreas_s> good question. If I understood ihar correctly, we want to try how it works out in Ocata
16:27:13 <andreas_s> and if we come to a good solution - yes
16:27:30 <andreas_s> more details probably in the summary mail
16:27:32 <andreas_s> #link http://lists.openstack.org/pipermail/openstack-dev/2016-November/106684.html
16:28:39 <rkukura> OK, lots of reading to catch up on
16:29:19 <rkukura> Anything else from the summit regarding ML2?
16:30:27 <andreas_s> not that I'm aware of
16:31:47 <rkukura> #topic Bugs/RFEs
16:32:05 <rkukura> I think we already covered port binding - anything more on that?
16:32:27 <rkukura> Or any other bugs/RFEs?
16:33:03 <rkukura> guess not…
16:33:15 <rkukura> #topic driver API feature requests
16:33:24 <rkukura> Anything anyone wants to discuss here?
16:34:18 <andreas_s> no
16:34:40 <rkukura> #topic Open Discussion
16:35:00 <rkukura> Anything else to discuss?
16:35:36 <rkukura> guess not
16:35:45 <yamamoto> hi
16:35:55 <yamamoto> just wanted to say hi. no topic.
16:35:57 <rkukura> hi yamamoto
16:36:25 <rkukura> are we done for today?
16:36:33 <yamamoto> i guess so
16:36:33 <rkukura> thanks for attending everyone!
16:36:42 <rkukura> #endmeeting