16:01:10 <Sukhdev> #startmeeting networking_ml2
16:01:10 <openstack> Meeting started Wed Mar 23 16:01:10 2016 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:14 <rkukura> hi
16:01:14 <openstack> The meeting name has been set to 'networking_ml2'
16:01:38 <Sukhdev> #topic: Agenda
16:01:46 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_March_23.2C_2016
16:02:09 <Sukhdev> Today's agenda is very brief
16:02:24 <Sukhdev> Hopefully we will finish early today
16:02:26 <yamahata> hello
16:02:44 <Sukhdev> #topic: Announcements
16:03:00 <Sukhdev> Anybody wants to make any announcement?
16:03:24 <Sukhdev> Mitaka is almost out and Newton is on its way
16:03:46 <yamamoto_> hi
16:04:18 <Sukhdev> #topic: Austin Summit Design summit Ideas
16:04:42 <Sukhdev> #link: https://etherpad.openstack.org/p/newton-neutron-summit-ideas
16:04:59 <Sukhdev> Have a look at the etherpad for design ideas
16:05:30 <Sukhdev> Please add ML2 related topics
16:06:56 <Sukhdev> rkukura : Do we want to add anything ML2 specific topic for discussion - perhaps we should discuss
16:07:14 <rkukura> Is there interest in tackling enforcement of extension semantics?
16:07:25 <rkukura> I’ve been thinking about adding that
16:08:22 <rkukura> And also possibly addiing support for MDs and EDs for all resources that are built into the core, including addres scopes, subnet pools, SGs, etc.
16:09:11 <Sukhdev> hmm... interesting
16:09:40 <Sukhdev> makes sense to me - what do others think?
16:11:19 <Sukhdev> rkukura : I am thinking another one - on the visibility of the dynamic segments created for HPB
16:11:20 <rkukura> I’m also wondering if some general discussion on the future of ML2 would be useful. The routed networks efforts is going to be making signficant changes, and things like segments that previously were completely internal to ML2 are getting more exposed in APIs and semantics, and also some of this seems to be moving from ML2 to core.
16:11:41 <rkukura> Sukhdev: I agree we should make them visibile, at least to admins
16:11:58 <rkukura> Do we need a discussion at summit on that? Do we have a bug report?
16:12:43 <Sukhdev> good question - we can address it either way -
16:13:24 <Sukhdev> rkukura : +1 on the general discussion on ML2
16:13:38 <rkukura> Maybe some general discussion on multi-segments would be helpful, covering what routed networks is trying to do with them, and exposing dynamic segments
16:14:04 <Sukhdev> correct - that is what I have in my mind
16:14:36 <Sukhdev> HPB presently talks about overlay networks - now, it is being addressed for routed networks -
16:14:49 <Sukhdev> so, it makes sense to discuss this
16:15:24 <rkukura> As part of that, I’d also like see if we want to reincarnate the old patches on making the network segment data model extensible (things other than phyiscal_network+segmentation_id)
16:16:18 <Sukhdev> yes - I like that
16:17:45 <rkukura> added it
16:18:30 <Sukhdev> cool - thanks - I was getting ready to assign you action item - you beat me to it :-):-)
16:19:26 <Sukhdev> Anything else on this?
16:20:02 <Sukhdev> #topic: Port Binding discussion related to migration
16:20:18 <Sukhdev> scheuran : are you here?
16:20:25 <scheuran> Sukhdev, yup
16:20:40 <scheuran> I'm currenlty working on a prototype
16:20:45 <Sukhdev> stage is yours
16:20:53 <Sukhdev> how is your finding?
16:21:05 <scheuran> I have moved the update binding:host_id into pre_live migration
16:21:15 <scheuran> it seems to work in my environment
16:21:23 <scheuran> I was able to migrate from a lb node to an ovs node
16:21:49 <scheuran> so technically this would be what I need
16:22:14 <rkukura> scheuran: That would be ideal if its all that immediately needed
16:22:18 <scheuran> but I think I will reach out via the mailing list to get more input about potential issues..
16:22:59 <Sukhdev> scheuran : I hope this does not break the main functionality
16:23:22 <scheuran> the logic of updating the xml seems to be very complex - I'm still working on it to cover all cases...
16:23:23 <rkukura> Do we think maintaining old and new bindings simultaneously (aka DVR distributed binding) will ever be needed?
16:23:54 <scheuran> Sukhdev, I haven't found something...
16:24:07 <Sukhdev> good
16:24:18 <scheuran> but I did not yet care about failure scenarios...
16:24:56 <Sukhdev> make sure have arosen review your patches - he is expert in that area
16:25:23 <scheuran> I'll do - I'll start with a mailing list reach out..
16:25:48 <Sukhdev> sounds like a good plan
16:26:38 <scheuran> nothing else on this topic from my side
16:26:42 <Sukhdev> anything else on this?
16:26:46 <scheuran> :)
16:27:25 <Sukhdev> typing fast/slow :-)
16:27:45 <Sukhdev> I kept the agenda light today -
16:27:53 <Sukhdev> #topic: Open Discussion
16:28:13 <Sukhdev> Anybody want to discuss anything general ML2 related
16:28:25 <Sukhdev> or we can tell each other some stories :-)
16:28:57 * Sukhdev waiting
16:30:13 <Sukhdev> Ok - I assume silence means no body has any thing to discuss
16:30:31 <rkukura> I’m on the west coast this week, so its early
16:31:03 <Sukhdev> ah ha rkukura is rubbing his eyes :-)
16:31:26 <rkukura> and burning my tongue on hotel coffee
16:31:43 <Sukhdev> :-)
16:32:19 <Sukhdev> good thing the time is changed otherwise, I usually run this meeting from my bed :-)
16:33:05 <Sukhdev> rkukura : If you are in town, perhaps get together for a beer one of these evenings?
16:33:22 <rkukura> Sukhdev: possibly tonight or tomorrow
16:33:57 <Sukhdev> sounds good - check your calender and ping me off-line
16:34:02 <rkukura> ok
16:34:24 <Sukhdev> OK - folks lets call it off before rkukura and I get into the brands of beers to try :-):-)
16:34:37 <Sukhdev> thanks for attending
16:34:41 <Sukhdev> bye
16:34:43 <rkukura> thanks Sukhdev!
16:34:43 <scheuran> thank you!
16:34:47 <scheuran> bye
16:34:50 <rkukura> bye
16:34:52 <yamamoto_> bye
16:34:54 <Sukhdev> #endmeeting