16:01:18 <Sukhdev> #startmeeting networking_ml2
16:01:18 <openstack> Meeting started Wed Nov 18 16:01:18 2015 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:22 <openstack> The meeting name has been set to 'networking_ml2'
16:01:34 <Sukhdev> #topic: Agenda
16:01:43 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_November_18.2C_2015
16:02:04 <Sukhdev> The agenda is continuation from the last week - hence, looks very similar
16:02:22 <Sukhdev> unless anybody wants to add anything?
16:02:43 <Sukhdev> #topic: Announcements
16:02:59 <Sukhdev> M1 is due in first week of Dec1
16:03:43 <Sukhdev> I hardly got settled from Japan trip and M1 is already knocking at the door :-):-)
16:04:11 <rkukura> related to this, I got the impression from this weeks neutron IRC meeting that we are close to having too many BPs for mitaka, so if anything needs a BP approved, get it in the pipeline!
16:04:49 <Sukhdev> yes, the list is very long
16:05:09 <Sukhdev> I missed the core meeting this week (too early for Pacific coasters)
16:05:23 <Sukhdev> anybody want to share anything interesting from that meeting?
16:06:22 <Sukhdev> #topic: Driver API for Security Groups
16:06:43 <Sukhdev> yamamoto is not here - I was looking for him earlier
16:07:05 <Sukhdev> I could help with this effort - but, I can reach him
16:07:13 <Sukhdev> anybody else is working on this?
16:07:26 <yamahata> yalei is working on it
16:07:35 <Sukhdev> This may require a BP, hence, the action need to take place soon
16:07:38 <yamahata> right now he's working for the BP
16:08:00 <yamahata> https://review.openstack.org/#/c/240356/
16:08:09 <Sukhdev> yamahata : cool - I will try to reach out to him - he is not here it seems
16:08:26 <yamahata> He's in china. It's midnight right now
16:09:05 <Sukhdev> I will try to reach to him later
16:09:13 <Sukhdev> anything else on this?
16:09:43 <Sukhdev> #topic: Driver status coalescing/visibility
16:10:18 <Sukhdev> yamahata: you and asoumya were looking to write a proposal on this
16:10:43 <Sukhdev> any update?
16:11:00 <yamahata> Sukhdev: do you mean yamamoto?
16:11:35 <Sukhdev> did I goof up on the name again - sorry :-)
16:11:57 <Sukhdev> asoumya is not here either
16:12:25 <yamahata> Ah no. You're correct. It's yamahata.
16:12:25 <Sukhdev> rkukura : are you aware of any work on this?
16:12:32 <yamahata> No update for it.
16:12:33 <rkukura> I’m not aware of any progress
16:12:39 <rkukura> This probably needs a BP
16:12:52 <Sukhdev> yamahata : ah ha - for a change I did not goof up :-):-)
16:13:43 <Sukhdev> #topic: Trunk port binding for VLAN-awere VMs
16:14:11 <Sukhdev> I do not know Bence's IRC - is he here?
16:14:48 <Sukhdev> I can help with this as well -
16:14:58 <rkukura> I was on the trunk port IRC meeting last night, but don’t recall his IRC name
16:15:26 <Sukhdev> This is needed for baremetal deployments as well -
16:16:07 <rkukura> I’m guessing he is rubasov, but not positive
16:16:46 <Sukhdev> I will sift through the blue print and patches to see if I can figure out and reach out to him :-)
16:17:18 <rkukura> This BP is being actively reviewed, and there will be impact on the OVS agent and on ML2’s port binding
16:17:32 <Sukhdev> correct
16:18:04 <Sukhdev> I will try to reach out to him -
16:18:13 <Sukhdev> #topic: Modular l2 agent & macvtap
16:18:27 <Sukhdev> scheuran : any update?
16:18:34 <rkukura> The nova cores have requested switching the trunk ports API to not adding a new resource
16:18:41 <scheuran> yes
16:18:49 <scheuran> I uploaded a first patchset
16:18:53 <scheuran> https://review.openstack.org/#/c/246318/
16:19:05 <scheuran> any input is helpful
16:19:27 <scheuran> this separates the linuxbridge impl specifics from the common agent loop stuff
16:20:16 <scheuran> and I had verified with the extension guys, that we won't collide during development
16:20:18 <Sukhdev> scheuran : cool - I will look at this later during the day
16:20:32 <scheuran> but we don't see any conflicts
16:20:39 <scheuran> Sukhdev, thanks
16:20:44 <scheuran> that's it
16:21:13 <rkukura> scheuran: Are you planning to file a BP for any of this, or just treat it as a bug fix?
16:21:38 <scheuran> rkukura, my plan is to treat it as bugfix
16:22:00 <scheuran> some folks from the drivers team were looking at it but none requested a spec
16:22:27 <rkukura> OK, as long as they are happy, I’m happy
16:23:12 <scheuran> rkukura, I mean it's basically what you find in the patchset I posted above + some extra testcases that I need to add
16:23:28 <scheuran> so it's nothing complex...
16:23:39 <rkukura> Sounds good - I’ll review it
16:23:44 <scheuran> rkukura, thanks!
16:23:51 <Sukhdev> scheuran : sounds good
16:24:05 <Sukhdev> anything else on this topic?
16:24:17 <scheuran> not from my side
16:24:23 <scheuran> will come back next week
16:24:33 <Sukhdev> cool - moving right along
16:25:04 <Sukhdev> I do not see shiv or arvind here, so I going to skip the Topology topic -
16:25:18 <Sukhdev> Unless somebody has any specific question on it?
16:25:54 <Sukhdev> #topic: Open Discussion
16:26:11 <Sukhdev> Anybody wants to discuss anything?
16:26:30 <Sukhdev> Looks like today's meeting is a short one :-)
16:27:21 <Sukhdev> rkukura : you mentioned something about nova people about the Nova and trunk port APIs -
16:28:30 <rkukura> Sukhdev: yes, the plan for trunk ports seems to be changing back to just extending port with an attribute to associate subports with it, rather than introducing a new trunkport resource.
16:29:09 <rkukura> With either API approach, there are lots of questions about how ML2 port binding will work with it.
16:29:20 <Sukhdev> oh - I remember that
16:29:42 <rkukura> For example, will binding:host_id and other portbinding attributes be applicable to the subports, or only to the master port.
16:29:43 <Sukhdev> this was one of the proposal at the summit discussion
16:30:05 <rkukura> And will the subports be bound individually or in buik
16:30:49 <Sukhdev> I do not believe you could bind them in bulk - as the sub-ports will be created on the fly
16:30:51 <rkukura> I think this gets kind of interesting with HPB - should the higher levels get bound individually, and then get tied together somehow at the lowest level
16:31:11 <rkukura> Sukhdev: Good point about dynamically adding/removing subports
16:31:50 <rkukura> One general question is how much detain on this needs to go in the BP vs. be worked out during implementation
16:32:33 <rkukura> How about if we try to get Bence and/or others to discuss the port binding for trunk ports with us at next week’s ML2 meeting?
16:32:43 <Sukhdev> Since this impacts multiple projects, I would like to see as much information as possible
16:33:11 <Sukhdev> rkukura : Good idea -
16:33:58 <Sukhdev> rkukura : I will try to look for him and you do too...
16:34:14 <Sukhdev> anything else on this or anything else?
16:34:21 <Sukhdev> otherwise, we are done....
16:34:34 <Sukhdev> thanks for attending -
16:34:37 <rkukura> I plan to try to attend future trunk port IRC meetings, and encourage other to as well
16:34:51 <rkukura> #link https://wiki.openstack.org/wiki/Neutron/TrunkPort
16:35:11 <rkukura> thanks Sukhdev!
16:35:25 <Sukhdev> rkukura : thanks for sharing the link
16:35:30 <yamamoto> rkukura: is it a regular meeting?
16:35:54 <rkukura> I think they are looking for a different time slot, and a channel with MeetBot
16:36:54 <Sukhdev> rkukura : once it is firmed up, please share the information with the team here
16:37:04 <rkukura> will do
16:37:21 <Sukhdev> Thanks folks - see you next week
16:37:24 <Sukhdev> bye
16:37:29 <yamamoto> bye
16:37:33 <Sukhdev> #endmeeting