16:02:02 <Sukhdev> #startmeeting networking_ml2
16:02:07 <openstack> Meeting started Wed Jul 22 16:02:02 2015 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:08 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:02:10 <openstack> The meeting name has been set to 'networking_ml2'
16:02:25 <Sukhdev> #topic: Agenda
16:02:30 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/ML2
16:02:43 <Sukhdev> Welcome to this week's ML2 meeting
16:03:04 <Sukhdev> #topic: Announcements
16:03:17 <Sukhdev> Liberty-2 is due next week
16:03:32 <Sukhdev> Do we have anything that we need to worry about?
16:03:50 <Sukhdev> #link: https://wiki.openstack.org/wiki/Liberty_Release_Schedule
16:04:13 <Sukhdev> Anybody has any other announcement?
16:04:19 * Sukhdev waiting
16:04:56 <Sukhdev> #topic: Action from previous weeks
16:05:11 <Sukhdev> ML2 mid-cycle plan
16:05:26 <Sukhdev> #link: https://etherpad.openstack.org/p/Neutron_ML2_Mid-Cycle_Sprint
16:05:57 <Sukhdev> manishg is not here
16:06:13 <rkukura> manishg did update with his availability
16:06:15 <Sukhdev> rkukura: have you had a chance to follow up with others on this?
16:06:31 <rkukura> I sent an email this morning to those who have expressed interest
16:06:58 <rkukura> myself, Sukhdev, and manishg are all available several of the possible weeks
16:07:14 <Sukhdev> kevinbenton: did you have a chance to look at the etherpad?
16:07:26 <Sukhdev> shivharris is not here either
16:07:34 <rkukura> I’d prefer the week of 9/28 - 10/2 since I can be in San Jose that entire week
16:08:15 <Sukhdev> rkukura: looks like that may be a good one -
16:08:29 <Sukhdev> lets see if others can sign up for that week as well
16:08:37 <rkukura> I’ll keep after Shiv and Josh, and post to openstack-dev as soon as we know what weeks are possible for those who have expressed interest so far.
16:09:07 <Sukhdev> sounds good
16:09:12 <Sukhdev> kevinbenton: are you here?
16:09:36 <rkukura> I’ll keep after kevinbenton as well
16:09:37 <Sukhdev> OK - lets move on
16:10:08 <Sukhdev> rkukura: cool - thanks - lets get this closed off - so that folks can get approvals and booking for travel
16:10:21 <rkukura> Sukhdev: +2000
16:10:53 <Sukhdev> #topic: L2 Modular Agent Discussion
16:11:15 <Sukhdev> banix does not seem to be here
16:11:40 <Sukhdev> anybody has any update on this topic?
16:11:49 <Sukhdev> otherwise, we will keep it on the agenda
16:11:54 * Sukhdev waiting
16:12:37 <Sukhdev> #topic: Physical Topology Discussion
16:13:15 <Sukhdev> asoumya or shivharris are not here
16:13:33 <Sukhdev> we'll skip this over as well
16:13:55 <Sukhdev> looks like we have lots of people on summer vacation
16:14:23 <Sukhdev> #topic: Ironic-ML2 discussion
16:14:42 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/Ironic-neutron
16:14:49 <Sukhdev> Let me provide an update on this
16:15:11 <Sukhdev> We have both of specs for Ironic-neutron integration approved
16:15:31 <Sukhdev> This means we are in the implementation phase
16:15:57 <Sukhdev> We have Ironic mid-cycle sprint in August (12-14), which I will be attending
16:17:00 <Sukhdev> I tested the end-to-end POC with fake data last week - just to make sure there is nothing missing from ML2 point of view
16:17:17 <Sukhdev> I was very impressed - all the plumbing works great
16:17:39 <Sukhdev> rkukura: congratulations on the ML2 framework -
16:18:13 <Sukhdev> This makes the implementation of this integration very smooth
16:18:23 <rkukura> Sukhdev: You are refering to the binding:profile and portbinding stuff that was done for SR-IOV, right?
16:18:44 <rkukura> And binding:vnic_type
16:18:49 <Sukhdev> rkukura: correct
16:19:52 <Sukhdev> I had thought that I will run into some holes and may have to submit additional patches - but, was impressed that everything is in place
16:20:27 <Sukhdev> anything else I can answer on this integration effort?
16:20:34 * Sukhdev waiting
16:20:54 <rkukura> Sukhdev: Sounds great!
16:21:02 <Sukhdev> cool -
16:21:17 <Sukhdev> We are at the end of our agenda -
16:21:41 <Sukhdev> since most of the folks are on vacation - I do not have anything additional to share
16:21:56 <Sukhdev> Shall we call it a day?
16:22:17 <rkukura> One item I’d like to mention
16:22:21 <Sukhdev> unless anybody who is in this meeting wants to bring up anything
16:22:28 <Sukhdev> rkukura: please do
16:22:55 <rkukura> I haven’t had a chance to dig into it yet, but the “routed network” effort seems to be moving forward, and could impact ML2
16:23:49 <rkukura> Some discussion is at https://bugs.launchpad.net/neutron/+bug/1458890
16:23:50 <openstack> Launchpad bug 1458890 in neutron "Add segment support to Neutron" [Undecided,Triaged]
16:23:51 <uvirtbot> Launchpad bug 1458890 in neutron "Add segment support to Neutron" [Undecided,Triaged]
16:23:52 <uvirtbot> Launchpad bug 1458890 in neutron "Add segment support to Neutron" [Undecided,Triaged] https://launchpad.net/bugs/1458890
16:24:08 <Sukhdev> rkukura: This was raised in the mid-cycle at Ft. Collins
16:24:10 <rkukura> And an email thread is at http://lists.openstack.org/pipermail/openstack-dev/2015-July/070028.html
16:24:35 <Sukhdev> there may be some impact - but, it is not very obvious on the surface
16:25:11 <rkukura> There was some discussion on whether these “segments” are related to ML2’s current notion of segments, but these seem to imply L2 bridging where ML2’s do.
16:26:02 <Sukhdev> rkukura: I had brought this up during mid-cycle
16:26:18 <rkukura> Anyway, just wanted to bring this to people’s attention again, and plan to try to undestand it myself
16:26:50 <rkukura> That’s all from me
16:26:57 <Sukhdev> so, ML2 segments could be used at the edge and in the middle will be L3 segments
16:27:28 <Sukhdev> rkukura: from the surface, it is like HPB - but, the middle is all L3 network
16:27:42 <rkukura> I’m not clear on this, but I think the idea is that tenant networks would not really be L2 domains
16:27:56 <Sukhdev> correct - that is the idea
16:28:12 <rkukura> So I’m not clear on how this fits the neutron model
16:28:16 <Sukhdev> there is an ehtherpad with some good use cases for this
16:28:17 <rkukura> Or can be made to fit
16:29:10 <Sukhdev> initially, we all had issue with how will it fit - but. then after lots of going back-and-forth and understanding use cases, this spec came about
16:29:29 <Sukhdev> I have not read it - I am going by what was discussed in the meeting
16:29:41 <Sukhdev> I will spend some timer to review this.
16:30:16 <rkukura> Sukhdev: I also am not clear on whether these new non-L2 segments are for users, or only for admins. If only for admins, there might be some overlap with the topology service being designed
16:31:02 <Sukhdev> rkukura: it is my understanding that this was all intended for admins - but, things may have changed from then to now :-)
16:31:51 <rkukura> We seem to be moving from neutron being focused on the user API, with plugins free to do whatever they want, to neutron exposing a consistent model to admins. I’m not sure if this is really a good thing or not.
16:32:29 <Sukhdev> rkukura: Please review the spec and comment on it -
16:32:48 <rkukura> Sukhdev: I will
16:32:55 <rkukura> Anything else?
16:33:02 <Sukhdev> nothing from me
16:33:15 <Sukhdev> Lets call it a day.
16:33:18 <Sukhdev> Thanks
16:33:28 <Sukhdev> bye
16:33:36 <Sukhdev> #endmeeting