09:00:03 <oanson> #startmeeting Dragonflow
09:00:04 <openstack> Meeting started Mon May 29 09:00:03 2017 UTC and is due to finish in 60 minutes.  The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:07 <openstack> The meeting name has been set to 'dragonflow'
09:00:17 <lihi> Hi
09:00:38 <dimak_> Hey
09:00:41 <oanson> Hi. Who's here for the weekly?
09:01:33 <oanson> #info lihi dimak_ in meeting
09:01:40 <oanson> itamaro, you're joining us?
09:01:53 <itamaro> Im In
09:02:00 <oanson> #info itamaro also in meeting
09:02:26 <oanson> nick-ma, hi
09:02:34 <nick-ma> hi
09:02:36 <oanson> #info nick-ma also in meeting
09:02:42 <oanson> All right. Let's get started.
09:02:47 <oanson> #topic Roadmap
09:03:08 <oanson> lihi just uploaded a new version of the portsec IPv6 patch. lihi - anything to report?
09:03:48 <lihi> I've added the Ping pong test. I'm hoping that the tests will succeed also on the gate :)
09:04:25 <oanson> portsec needs a ping-pong test?
09:05:02 <lihi> Yes, from different ip and mac
09:05:23 <oanson> I see. All right.
09:05:41 <oanson> SFC - dimak_ I saw you uploaded many patches. Sorry I didn't review them yet
09:05:52 <dimak_> Yeah
09:06:05 <dimak_> I still have to address some comments
09:06:16 <oanson> Sure
09:06:17 <dimak_> I've managed to rebase it on top of lport
09:06:32 <oanson> That's good. I understood that was blocking
09:06:46 <dimak_> and made it work both for mpls and non-mpls SFs
09:06:51 <dimak_> yeah
09:06:53 <oanson> Cool!
09:07:05 <dimak_> And I still need to fix some things for multinode env
09:07:17 <dimak_> But on a single node all my tests passed
09:07:32 <dimak_> I hope we can get them to pass in CI as well
09:07:34 <oanson> All right. Multi-node can be a phase 2.
09:07:49 <dimak_> This should help https://review.openstack.org/#/c/467259/
09:07:53 <oanson> As in - get the patches that are live now to work on single node, and then add support for multinode.
09:08:18 <oanson> Yes - it already has 2 +2s, so I guess it will make it in soon
09:08:27 <dimak_> Not sure why we need neutron liasion
09:08:36 <dimak_> liaison
09:08:59 <oanson> Dunno either. But it shouldn't be a problem. Do you know who the liaison is?
09:09:04 <dimak_> I'll drop by infra channel to check it out
09:09:18 <oanson> Thanks
09:09:34 <dimak_> Thats all on SFC end
09:09:50 <oanson> Northbound API - there are a couple of patches left. And I have to do the migration for the lport migration model
09:10:00 <oanson> I think that would be it - we can start working on cleanup
09:10:35 <dimak_> Active ports
09:10:38 <oanson> And active ports. Yes
09:10:44 <dimak_> I have patches for OvsPort and FIP up for review
09:10:46 <oanson> (I think that one's also on me)
09:11:15 <oanson> I'll give the active port and migration priority on my end
09:11:17 <dimak_> I can do that if your hands are full
09:11:31 <oanson> Nah. I'm good.
09:11:51 <oanson> If you're free I have other things we can work on :)
09:11:59 <oanson> We'll get to that in a minute.
09:12:13 <oanson> Lastly - VLAN aware VMs is up for review, and passes gate and manual tests.
09:12:24 <lihi> 🎉
09:12:28 <oanson> Works automatically if q-trunk service is enabled in local.conf
09:12:50 <nick-ma> great job, i'm reviewing.
09:12:51 <dimak_> I'll review it today
09:12:57 <oanson> Thanks
09:13:05 <oanson> Since all our goals are so close to being finished, I want to discuss future plans:
09:13:37 <oanson> 1. Stabalising OSA Deployment
09:14:00 <oanson> It was a bit abandoned, and I want to get back to it before we loose momentum.
09:14:35 <oanson> As part of that, and seeing that etcd is now a base service, I think having an etcd based publisher is also a good idea
09:14:44 <oanson> lihi, you said you'll take this?
09:14:49 <nick-ma> maybe we can set up an etherpad for discussion.
09:14:54 <oanson> Sure
09:14:56 <dimak_> I plan on doing DF skydive plugin
09:15:01 <lihi> Yes, I'm looking into it
09:15:19 <oanson> #link Dragonflow Pike Roadmap https://etherpad.openstack.org/p/dragonflow-pike
09:15:29 <oanson> It already exists....
09:17:38 <oanson> I also added LBaaS
09:17:52 <oanson> It doesn't have an owner, so I'll take it for now.
09:18:06 <oanson> Unless we have a volunteer
09:18:28 <oanson> I think it will be good to collaborate with Kuryr and Octavia. That's why I want to push it forwards.
09:19:08 <oanson> dimak_, I see that RPM packaging is here, and I remember you volunteered for it before. Do you still want it?
09:19:29 <dimak_> Yeah, I began looking into it
09:19:45 <dimak_> But that work got stalled by other things :(
09:20:01 <oanson> Yes. No complaints. Just trying to organize the effort
09:20:47 <oanson> All right. I think this fills our feature buffer.
09:21:02 <oanson> Anything else for roadmap?
09:21:55 <oanson> All right.
09:21:58 <oanson> #topic Bugs
09:22:40 <oanson> Regarding bug #1647362
09:22:41 <openstack> bug 1647362 in DragonFlow "Fullstack tests does not simulate actual networks settings" [High,In progress] https://launchpad.net/bugs/1647362 - Assigned to Lihi Wishnitzer (lihiwish)
09:23:01 <oanson> Change https://review.openstack.org/#/c/467251/ states it's a partial solution.
09:23:11 <oanson> What's missing?
09:23:51 <oanson> lihi? It's your patch
09:23:56 <lihi> Local port is assigned to a subnet, and not to a network and The router should be connected to a network, and to a subnet
09:24:19 <oanson> I see. All right. That's a shame.
09:24:33 <oanson> Bug #1664967
09:24:35 <openstack> bug 1664967 in DragonFlow "SGApp: unsupported operand type(s) for +: 'int' and 'NoneType'" [High,Incomplete] https://launchpad.net/bugs/1664967 - Assigned to Lihi Wishnitzer (lihiwish)
09:25:07 <oanson> nick-ma says it's no longer valid - the exception no longer appears.
09:25:15 <oanson> From a short look I took, I agree.
09:25:18 <oanson> Can it be closed?
09:25:36 <lihi> Yes, I think so
09:25:42 <oanson> Would you mind?
09:25:47 <lihi> Sure
09:26:17 <oanson> Thanks
09:26:41 <oanson> Bug #1690775
09:26:42 <openstack> bug 1690775 in DragonFlow "Remove special handling for lport/ofport in local controller" [High,New] https://launchpad.net/bugs/1690775 - Assigned to Omer Anson (omer-anson)
09:26:50 <oanson> This bug bugs me, so I started hacking at it
09:27:22 <oanson> I plan  to make many small changes until it is fixed. Hopefully, it will be many byte-sized changes that won't be too hard to review
09:27:36 <oanson> I don't want another 1000-line patch with my name on it :)
09:27:45 <oanson> Anything else for bugs?
09:28:11 <oanson> #topic Open Discussion
09:28:29 <oanson> I forgot to bring this up in roadmap - but it would be nice to offer Dragonflow as an L3 flavor
09:28:35 <oanson> This was recommended by the Neutron folks.
09:28:57 <oanson> Is anyone free enough to take this?
09:29:09 <dimak_> Maybe we should draft up some spec for this
09:29:19 <oanson> Sounds good.
09:29:31 <oanson> We told Neutron folks it probably won't be ready this cycle. But it would be nice if we made it early.
09:29:38 <oanson> dimak_, do you have the time for it?
09:29:39 <dimak_> Seeing that DF's architecture does not allow this out of the box
09:29:50 <dimak_> I'll start looking into it
09:30:21 <oanson> That's not good news. The idea is pluggability - that we can run with only some apps features and it would work
09:30:38 <oanson> In the spec, it might be helpful to outline where that assumption breaks
09:30:57 <dimak_> I'll look into it and update next weekly
09:31:07 <oanson> Great. Thanks!
09:31:14 <oanson> I am releasing the floor to the wil
09:31:16 <oanson> wild*
09:31:41 <oanson> Anyone has anything to bring up?
09:31:46 <dimak_> We won't have a meeting next week, right?
09:31:54 <oanson> Why not?
09:31:57 <dimak_> There's Openstack Day IL
09:32:07 <dimak_> I think most of us will be there
09:32:24 <oanson> Yes. That's true.
09:33:07 <oanson> All right. Then yes, next week's meeting is canceled. We'll meet again on the 12th of June
09:33:30 <oanson> Anything else?
09:34:04 <oanson> All right. Thanks everyone for coming.
09:34:14 <oanson> #endmeeting