09:00:15 <oanson> #startmeeting Dragonflow
09:00:16 <openstack> Meeting started Mon Jan 30 09:00:15 2017 UTC and is due to finish in 60 minutes.  The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:19 <openstack> The meeting name has been set to 'dragonflow'
09:00:23 <oanson> Hi. Who's here for the Dragonflow weekly?
09:00:44 <lihi> Hi
09:00:59 <oanson> If it's just the two of us, we have an issue.
09:01:15 <oanson> I understood it's the new year in china, so they are in holiday
09:01:48 <dimak> Hey
09:01:52 <oanson> Let's wait another minute, and take it from there
09:02:40 <irenab> hi
09:02:45 <yuli_s> hello
09:03:17 <oanson> Since we are a skeleton crew today, I'll just run through the stuff real quick
09:03:33 <ishafran> Hi
09:03:37 <oanson> IPv6 - Waiting on lihi for security group support.
09:03:41 <oanson> SFC - no change.
09:03:58 <oanson> Service health reporting - no change (AFAIK)
09:04:05 <oanson> Tap as a service - No change (AFAIK)
09:04:15 <oanson> sNAT - There is a new patch set. Please review
09:04:26 <lihi> I'll upload a patch today for the sg
09:04:26 <oanson> NB Refactor - There are new patches. Please review.
09:04:27 <itamaro> Hi All
09:04:40 <oanson> dimak, note that the tests are failing on the more advanced stuff. Please fix.
09:05:16 <oanson> Native API - There is a spec, but it's WF-1. I think it is ready for review/voting
09:05:18 <dimak> Yeah, openstack picked a new version of jsonmodels and I need to fix something there, I hope it gets merges and released by tomorrow
09:05:28 <dimak> *openstack-requirements
09:05:41 <oanson> dimak, sound good, oh jsonmodels lead commiter!
09:06:03 <irenab> oanson, I added working items to the API spec, will upload the patch in few mins
09:06:16 <oanson> Great! Once that's done, it can be reviewed
09:06:20 <yuli_s> i am trying to push my commits to be finally released.
09:06:30 <oanson> yuli_s, great!
09:06:41 <oanson> Anything else for roadmap?
09:07:08 <oanson> #topic bugs
09:07:19 <dimak> I hope we can merge nb refactor spec too :)
09:07:31 <oanson> Hopefully
09:07:35 <dimak> I'll remove the line Irena pointed out if there's another PS
09:07:51 <oanson> nick-ma added a spec for the duplicate publisher issues: https://review.openstack.org/#/c/425619/
09:08:14 <oanson> I'd be happy if it can be reviewed. It looks very interesting
09:08:21 <oanson> That's all I have for bugs.
09:08:37 <oanson> Anything else, or can we move on?
09:08:41 <yuli_s> yes
09:08:57 <oanson> yuli_s, shoot!
09:09:00 <yuli_s> an old problem with private network tenant is getting back
09:09:18 <yuli_s> I need to do "NEUTRON_CREATE_INITIAL_NETWORKS=False"
09:09:23 <yuli_s> to run stack.sh
09:09:53 <oanson> yuli_s, I ran stack yesterday, and it worked
09:10:02 <oanson> Are you sure it's not an environment issue?
09:10:07 <yuli_s> hm, you have this line in configuration ?
09:10:36 <yuli_s> by default it should be true
09:10:38 <oanson> Nope
09:10:46 <oanson> No such line in my local.conf
09:10:46 <lihi> I think I ran into this issue also
09:11:05 <lihi> On new VM and env
09:11:11 <oanson> lihi, can you verify? Does adding this line help?
09:11:40 <lihi> I'll try. I did not knew adding it might help
09:11:59 <oanson> Anyone else seeing this issue?
09:12:33 <oanson> All right. lihi, yuli_s, let's take this offline.
09:12:59 <oanson> Anything else for bugs?
09:13:37 <oanson> #topic Open Discussion
09:13:44 <oanson> The floor is for the taking
09:13:58 <itamaro> There is a small bug I encountered while doing my tests, I published a fix,Pls review it
09:14:19 <oanson> itamaro, what's the link?
09:14:24 <itamaro> https://review.openstack.org/#/c/426601/
09:14:51 <dimak> Will take a look
09:15:08 <oanson> Same here
09:15:10 <itamaro> :)
09:15:16 <oanson> The floor is free once more
09:15:45 <dimak> I have a patch for datapath prop
09:15:56 <dimak> I want it in quite fast as it is a pain to rebase
09:16:11 <dimak> https://review.openstack.org/#/c/426584/
09:16:46 <oanson> Hmm....
09:17:21 <itamaro> It has a lot of lines. but the change is stright forward :)
09:17:30 <oanson> dimak, any chance we give nick-ma/xiaohhui another day to look at it?
09:17:35 <dimak> Sure
09:17:37 <oanson> Yes
09:17:50 <oanson> yuli_s, please review it too.
09:18:15 <dimak> If nick-ma is not around then we don't have 2x +2 for other patches as well :P
09:18:33 <oanson> Yes. But he will be back next week.
09:18:46 <oanson> All right. Let's wait until tomorrow.
09:18:49 <dimak> I meant that it can wait without rebasing
09:18:59 <dimak> Because other patches will wait too
09:19:28 <oanson> yuli_s, please review it as well. I will feel better pushing it if your review is there as well
09:19:55 <oanson> Other patches can already rebase on top of this patch. I doubt it will have any major changes, even if there are negative reviews.
09:20:14 <oanson> And I agree it's important
09:20:21 <oanson> Makes our code simpler
09:20:44 <oanson> ishafran, if you want to review it as well, that can also help
09:20:51 <oanson> Basically, I want another review! :)
09:21:02 <ishafran> OK I'll review it too
09:21:06 <oanson> Thanks!
09:21:11 <oanson> The floor is for the taking again
09:21:19 <ishafran> I reviewed changes required for different modes of SNAT: global IP/tenant based IP. Current intention to separate SNAT application into two different applications with common code base
09:21:40 <ishafran> Are we OK with it?
09:22:26 <oanson> I am, but I am not sure everyone knows what this is about. Maybe give some context?
09:22:50 <itamaro> Why no do it in one apllication and have  cofiguration dictate behevior
09:22:54 <ishafran> it described well in spec
09:22:56 <itamaro> not
09:23:52 <ishafran> I tried to make it this way - too much ifs
09:24:50 <itamaro> it can be a compiste class or impl-pimpl kind of implemantation
09:24:50 <oanson> Fine. I'll give the context :)
09:25:04 <oanson> The snat application is designed to have two modes of operation
09:25:14 <oanson> 1. single external IP assigned to the host/compute node
09:25:24 <oanson> 2. A pool of IPs, each assigned to <host, tenant> pair
09:25:33 <oanson> no. 1 is up for review
09:26:01 <oanson> no. 2 is now being written. Will it be written as part of the original application (option 1), or as a second application?
09:26:08 <oanson> And extract a base class to parent both apps.
09:26:15 <oanson> ishafran, is this summary correct?
09:26:18 <itamaro> understood.
09:26:37 <ishafran> summary is perfect :)
09:27:13 <itamaro> still think an envolope class that 'loads' the right impl is a ggod way to do it.
09:27:46 <ishafran> you mean dynamically loaded implementation
09:27:47 <oanson> itamaro, that's an infrastructure we don't have, that will be used only for one app
09:28:09 <oanson> We already have such a configuration - it states which apps to load :)
09:28:31 <itamaro> OK
09:28:50 <oanson> All right, so we're agreed - 2 apps, with a parent class?
09:29:15 <dimak> Or a common module they use :)
09:29:26 <ishafran> I guess I will implement it separately and then I'll think how to pack it together
09:29:31 <oanson> dimak, even better, if it can be done
09:29:42 <oanson> ishafran, sounds like a plan
09:30:22 <oanson> Are we all agreed then?
09:30:32 <itamaro> V
09:30:41 <oanson> The floor is free once more
09:31:01 <oanson> The floor is lava - take it while it's hot!
09:31:29 <oanson> All right. Thanks everyone for coming.
09:31:40 <oanson> #endmeeting