09:00:16 <oanson> #startmeeting Dragonflow
09:00:17 <openstack> Meeting started Mon Nov 28 09:00:16 2016 UTC and is due to finish in 60 minutes.  The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:18 <dimak> Hello
09:00:20 <openstack> The meeting name has been set to 'dragonflow'
09:00:28 <oanson> Hi.
09:00:33 <oanson> Who's here for the Dragonflow meeting?
09:00:35 <ntr0py> Hi
09:00:37 <lihi> Hi
09:00:42 <xiaohhui> hello
09:02:07 <oanson> All right, let's get started
09:02:22 <oanson> We'll use the same agenda as last week.
09:02:44 <oanson> It's almost the same agenda, and I didn't have a chance to update it (Had a meeting this morning)
09:02:52 <irenab> hi
09:03:10 <oanson> #info dimak ntr0py liangy xiaohhui irenab in meeting
09:03:24 <oanson> #info lihi is also in meeting
09:03:28 <oanson> #topic Ocata roadmap
09:03:42 <oanson> #link SFC spec https://review.openstack.org/#/c/394498/
09:03:55 <oanson> This spec looks almost complete. I will pass is once this meeting is over
09:03:56 <yuli_s> Hello
09:04:02 <oanson> I didn't have a chance to do that earlier.
09:04:16 <dimak> Thanks, I hope it will get merged today then
09:04:22 <oanson> #link Chassis is alive spec https://review.openstack.org/#/c/385719/
09:04:30 <oanson> This spec is merged.
09:04:51 <oanson> I understand rajivk will also provide a follow up spec.
09:05:08 <rajivk> I already provided :)
09:05:11 <oanson> rajivk, do you want to comment, or are you keeping the suspense until the spec is up?
09:05:19 <oanson> Great!
09:05:41 <rajivk> #link https://review.openstack.org/#/c/402395/
09:05:51 <oanson> Beat me to it :)
09:06:10 <rajivk> Please provide your views :)
09:06:21 <rajivk> On specs
09:06:36 <oanson> I see nick-ma_ already put some comments. I'll go over it the first chance I get
09:06:47 <nick-ma_> hi
09:06:59 <oanson> #info nick-ma_ also in meeting
09:07:01 <oanson> nick-ma_, hi
09:07:17 <oanson> #link TaaS spec https://review.openstack.org/#/c/396307/
09:07:40 <oanson> This spec is also very far along. I think the only thing left is the 'position' argument.
09:07:44 <oanson> yuli_s, am I correct?
09:07:44 <yuli_s> I was on vocation for a few days
09:07:54 <yuli_s> so, had no chance to finish it
09:08:02 <yuli_s> hope to send an updated version today
09:08:03 <oanson> yuli_s, no worries. :)
09:08:10 <oanson> No shortage of work for reviewers :)
09:08:24 <yuli_s> ok ;)
09:08:31 <oanson> Looking forward to seeing you update.
09:08:41 <yuli_s> ok
09:09:02 <oanson> #link anonymous sNAT spec https://review.openstack.org/#/c/397992/
09:09:35 <oanson> Igor, are you here?
09:09:52 <oanson> I'll hunt him down offline.
09:10:02 <oanson> In any case, it looks like there is still a lot of work here.
09:10:14 <oanson> Anything specific that has to be discussed?
09:10:53 <nick-ma_> i noticed the northbound api, any specific plan for that part?
09:11:09 <oanson> I think Igor is working on adding it to Neutron
09:11:16 <oanson> Let me see if I can find his blueprint
09:11:48 <nick-ma_> ok. it is a huge work. i'm looking forward to the object model.
09:12:02 <irenab> oanson: was this agreedto be accepted by neutron drivers?
09:12:19 <oanson> irenab, As far as I know, it's still in the pipes
09:12:22 <ishafran> Hi, I am back
09:12:22 <nick-ma_> why need agreement of neutron drivers?
09:12:43 <oanson> I can't find it in the blueprints. I'll have to ask him.
09:13:02 <irenab> if the change to neutron, it should be agreedby the team
09:13:15 <irenab> if its only DF,then it does not
09:13:16 <nick-ma_> i cannot see any changes to neutron.
09:13:23 <ishafran> Here is a blueprint but it requires a change: https://blueprints.launchpad.net/neutron/+spec/neutron-local-snat
09:14:02 <oanson> #link Distributed gateway for optimized local SNAT (Neutron Blueprint) https://blueprints.launchpad.net/neutron/+spec/neutron-local-snat
09:14:27 <nick-ma_> ok, got it.
09:14:30 <oanson> ishafran, what's the status of it in Neutron? I can't see any discussion?
09:14:49 <ishafran> I have a reference implementation based on mitaka
09:14:52 <nick-ma_> i just notice that we are in the discussion of snat.
09:15:07 <ishafran> RFE itself is confirmed and in wishlist
09:15:26 <oanson> nick-ma_, what topic were you asking about?
09:15:39 <irenab> ishafran: seems its good to go on then
09:15:58 <irenab> not before Pika though
09:16:15 <ishafran> irenab: yes I guess
09:16:33 <oanson> ishafran, we can add the API support in Pike, but the anonymous snat feature as a DF app in Ocata
09:16:50 <oanson> control it via configuration in Ocata, until API is merged (or at least in review)
09:17:36 <oanson> Can we move on?
09:18:06 <ishafran> oanson: we can delay an API, in general DF support for this feature is much more flexible
09:18:44 <oanson> ishafran, I think we're in agreement.
09:18:52 <oanson> Once the API exists, we can add support for it
09:19:05 <oanson> #link LBaaS spec #link LBaaS spec https://review.openstack.org/#/c/397997/
09:19:31 <oanson> About LBaaS, I'll hunt down denghui and ask him to update the spec
09:19:50 <oanson> I'll also ask him to loiter around in the dragonflow channel in case anyone wants to strike up a discussion
09:20:30 <oanson> #link Dragonflow deployment in openstack-ansible https://review.openstack.org/#/c/391524/
09:21:11 <oanson> This is coming along slowly but surely. It looks like the base is in place. There are a couple of things I want to add to the testing (output df-local-controller log, output subunit results to file)
09:21:24 <oanson> Once that is done, I will ask the OSA guys if we can start the review process.
09:21:51 <oanson> There is a lot of work still on Dragonflow side, but I think the OSA side is (alomost) good to go (except metadata service, which I'll add later)
09:22:22 <oanson> Since metadata service requires a bit of network tweaking which is very devstack specific currently, and I want to see if it can be made more general
09:22:30 <oanson> Or if the way it works currently is good enough.
09:22:44 <oanson> Anything else for roadmap?
09:23:36 <oanson> #topic Bugs
09:24:14 <oanson> It looks like most of the bugs have owners.
09:25:09 <oanson> There is one critical bug by xiaohhui
09:25:29 <oanson> xiaohhui, I am trying to understand the reviews for this bug.
09:25:43 <oanson> Is it https://review.openstack.org/#/c/385350/ and https://review.openstack.org/#/c/378192 ?
09:26:13 <xiaohhui> what bug are you referencing?
09:26:43 <oanson> Sorry, this bug: https://bugs.launchpad.net/dragonflow/+bug/1628737
09:26:43 <openstack> Launchpad bug 1628737 in DragonFlow "Exception occurs when removing last port of compute node" [Critical,In progress] - Assigned to Hong Hui Xiao (xiaohhui)
09:27:44 <oanson> I see you updated the duplicate bug 1641903 with the reviews
09:27:44 <openstack> bug 1628737 in DragonFlow "duplicate for #1641903 Exception occurs when removing last port of compute node" [Critical,In progress] https://launchpad.net/bugs/1628737 - Assigned to Hong Hui Xiao (xiaohhui)
09:28:13 <oanson> If the last comment is accurate, I can copy the comment to the original bug
09:28:26 <oanson> He left :(
09:28:49 <oanson> Probably not on purpose.
09:29:09 <oanson> Anyway, it looks like the last comment on https://bugs.launchpad.net/dragonflow/+bug/1641903 states the current status. I'll update the original bug
09:29:09 <openstack> Launchpad bug 1628737 in DragonFlow "duplicate for #1641903 Exception occurs when removing last port of compute node" [Critical,In progress] - Assigned to Hong Hui Xiao (xiaohhui)
09:29:16 <oanson> Anything else in bugs?
09:29:40 <oanson> #topic Open Discussion
09:29:44 <oanson> The floor is for the takinhg
09:29:50 <oanson> taking*
09:29:56 <nick-ma_> cassandra driver is working.
09:30:58 <rajivk> I have suggestion to everyone, the code contains very large methods. If everyone from now own takes care of writing small methods then it might be readable.
09:31:12 <oanson> nick-ma_, that's great. Looking forwards to trying it out.
09:31:33 <oanson> rajivk, that's a good suggestion.
09:31:43 <nick-ma_> we also have several pending BIG reviews, which were proposed in newton.
09:31:58 <oanson> Personally I prefer methods no longer than 20 lines
09:32:15 <oanson> nick-ma_, yes. They also conflict with one another.
09:32:16 <nick-ma_> like port-qos, topology data consistency, allow address pair, etc.
09:32:28 <rajivk> i too. May be core reviewers can make sure. :)
09:32:46 <oanson> I'll try to tackle one a day, in this order: DB consistency, active-notify app, qos
09:32:50 <xiaohhui> my connection is interrupted
09:33:01 <oanson> xiaohhui, no worries.
09:33:07 <xiaohhui> The fix will be https://review.openstack.org/#/c/385350/7
09:33:23 <oanson> xiaohhui, I said I'll take the last comment on bug 1641903 and copy it to the original bug
09:33:23 <openstack> bug 1628737 in DragonFlow "duplicate for #1641903 Exception occurs when removing last port of compute node" [Critical,In progress] https://launchpad.net/bugs/1628737 - Assigned to Hong Hui Xiao (xiaohhui)
09:33:23 <xiaohhui> however, there is a dependent patch https://review.openstack.org/#/c/385349/10
09:33:34 <oanson> If anyone prefers a different order, please let me know.
09:33:45 <oanson> And I apologize reviews on my end take so long.
09:34:44 <nick-ma_> oanson: take it easy.
09:34:59 <rajivk> oanson, no worries :)
09:35:13 <oanson> :)
09:35:24 <oanson> Any other open items?
09:36:08 <oanson> All right. Thanks everyone for coming.
09:36:14 <nick-ma_> ok
09:36:16 <nick-ma_> thanks.
09:36:23 <xiaohhui> dinner time
09:36:25 <oanson> I think the project is really moving forwards! Thanks for your efforts!
09:36:30 <oanson> Bon appetit!
09:36:30 <irenab> thanks
09:36:39 <oanson> #endmeeting