09:01:35 <oanson> #startmeeting Dragonflow
09:01:36 <openstack> Meeting started Mon Jan  9 09:01:35 2017 UTC and is due to finish in 60 minutes.  The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:01:37 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:01:39 <openstack> The meeting name has been set to 'dragonflow'
09:01:47 <oanson> Who is here for the Dragonflow meeting?
09:01:51 * xiaohhui raises his hand
09:01:57 <dimak> Good morning
09:02:02 <oanson> yuli_s, nick-ma_, lihi ?
09:02:27 <oanson> irenab ?
09:02:31 <oanson> All right. Let's start
09:02:39 <oanson> #info dimak and xiaohhui in meeting.
09:02:39 <irenab> hi
09:02:46 <oanson> #info irenab is in meeting too
09:03:02 <oanson> #topic Roadmap
09:03:40 <oanson> Since we are so few hands today, I'll just run through what I know about the roadmap, and you'll add anything I will missed.
09:04:10 <oanson> IPv6 - I know lihi is working on the comments to the neighbour discovery / advertising patch.
09:04:27 <oanson> #link Add IPv6 ND Neighbor Advertiser application https://review.openstack.org/#/c/412208/
09:05:12 <oanson> SFC had no progress due to the work on NB refactor
09:05:31 <oanson> Chassis health and service health, rajivk uploaded a patch with implementation
09:05:35 <dimak> 😞
09:05:49 <oanson> #link Added support for service status reporting https://review.openstack.org/#/c/415997/
09:06:09 <oanson> I think we should have the openstack bot quote patches in the same way it quotes bugs
09:06:52 <oanson> I have to go over the specs again, but rajivk's patch definitely includes the second spec (Spec to support service status reporting https://review.openstack.org/#/c/402395/ )
09:07:07 <xiaohhui> About chassis health, I will put my spirit in review and code base on rajivk's work.
09:07:24 <oanson> xiaohhui, great. Thanks! I think he would appreciate that too.
09:07:33 <oanson> As far as I know, there is no progress in TAPaaS.
09:07:54 <oanson> yuli_s has been working on aggregating the rally results, and didn't have a lot of time to look into it
09:08:22 <oanson> NB API - We are still trying to get infra to accept jsonmodels. They are trying to convince us to use oslo.versionedobjects (OVO)
09:08:38 <dimak> The needed fixes were merged
09:08:49 <oanson> We think that OVO is a bit of an overkill for what we want to do, whereas jsonmodels is very simple.
09:08:56 <oanson> dimak, great. Thanks!
09:09:07 <dimak> And I honestly think we're not at the point yet where we can benefit from OVO
09:09:42 <oanson> #link Add jsonmodels requirement https://review.openstack.org/#/c/415164/
09:09:58 <oanson> I see Matthew Thode removed his objections.
09:10:15 <dimak> I'm trying to get tonyb on IRC but timezone differences make it a bit difficult
09:11:07 <oanson> All right.
09:11:27 <oanson> Anything else for roadmap?
09:11:31 <dimak> I'll upload some refactor patches
09:11:52 <dimak> I've started to port Chassis model so we can see how it will look
09:12:06 <oanson> dimak, Great.
09:12:17 <dimak> I'd love some input whether it makes things better or worse before we port other models :)
09:12:20 <oanson> I'll also update the spec today to incorporate everyone's latest comments.
09:12:38 <oanson> #topic Bugs
09:13:01 <oanson> I don't see anything new and urgent except one issue
09:13:09 <oanson> Bug 1651643
09:13:09 <openstack> bug 1651643 in DragonFlow "metadata service cannot start due to zmq binding conflict" [Critical,In progress] https://launchpad.net/bugs/1651643 - Assigned to Li Ma (nick-ma-z)
09:13:43 <oanson> It's popping up all the time, forcing services to fail since we have too many publishers/subscribers.
09:14:23 <oanson> I see nick-ma_ is on it. But he isn't here now.
09:14:30 <xiaohhui> What is the specific output of the error?
09:14:43 <oanson> It's in the bug report:
09:14:49 <xiaohhui> I see
09:14:50 <oanson> ZMQError: Address already in use
09:15:04 <dimak> I see address in use in df-local-controller occasionally too
09:15:12 <oanson> Yes
09:15:47 <oanson> I think the issue is that we were lazy, and thought that in compute node side, we don't need the publisher service for multiproc publishers
09:15:49 <oanson> (e.g. zmq)
09:16:29 <oanson> The truth is I'm not sure we need a publisher on the metadata service either. I think that's an error since everyone uses nb_api.
09:17:52 <oanson> Any other bugs that need to be discussed?
09:18:33 <oanson> #topic Open Discussion
09:18:44 <oanson> The floor is for the taking.
09:19:03 <dimak> Nothing on my end
09:19:20 <xiaohhui> Nothing from me too
09:19:34 <oanson> All right. Then thanks everyone!
09:19:42 <oanson> Keep up the great work!
09:19:45 <dimak> Good bye!
09:19:48 <oanson> #endmeeting