19:00:35 #startmeeting Dragonflow 19:00:36 Meeting started Mon Dec 18 19:00:35 2017 UTC and is due to finish in 60 minutes. The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:37 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:00:39 The meeting name has been set to 'dragonflow' 19:00:43 o/ 19:00:50 o/ 19:01:12 CrayZee, you with us too? 19:01:15 yup 19:01:23 #info dimak leyal CrayZee are in meeting 19:01:40 All right. I think that's the best we're gonna get, and better than usual for this time :) 19:01:45 #topic Roadmap 19:01:57 No updates on LBaaS (As far as I know) 19:02:13 None from my side at least 19:02:13 lihi started looking at DNS integration with Dragonflow. I think there's also a spec. 19:02:45 https://review.openstack.org/527956 19:02:55 CrayZee, beat me to it :) 19:03:03 #link DNS as a service spec https://review.openstack.org/#/c/527956/ 19:03:05 She offered two options and asked to vote 19:03:21 also bit me :) 19:03:43 leyal, please tell me you meant beat and not bit :) 19:03:48 :D 19:04:03 All right. So dimak and I will have to review it and vote. 19:04:11 Adding myself 19:04:12 :D 19:04:27 Upgrades 19:04:36 We discussed it earlier 19:04:52 to support migration code first 19:05:03 or independently, if not first 19:05:04 We had a discussion offline this morning - if I recall, we agreed that we have a migration script for Dragonflow, an Alembic Migration for Neutron DB,? 19:05:25 dimak, not sure I understand what you meant? 19:05:29 I think we just need DF upgrages 19:05:34 and DF migration 19:05:40 that will also reconcile existing data 19:05:55 But some changes need to be done in Neutron DB too 19:06:09 The DHCP ports is an excellent example 19:06:20 Can we make those online? 19:06:47 I think we kind of can 19:07:04 I'd prefer not, for two reasons: 1. We'll be dependent on Neutron code for upgrade. 2. We might end up needing both versions of the model 19:08:21 Umm 19:08:22 dimak , i think that will cause to events on objects that already exsist in DF db 19:08:30 Why? 19:08:52 Let's go about this differently - let's outline the 2-3 options we have in the spec, and take the discussion there. I don't want to eat up too much meeting time for this. 19:08:53 we add those object during the DF-DB upgrade , and then create them in neutron .. 19:09:02 dimak, if you could add your suggestion? And I will add mine? 19:09:08 Sure 19:09:14 Great. Thanks! 19:09:33 #action oanson dimak update upgrade spec 19:09:44 Deployment 19:10:08 I understand lihi is making some progress with OSA, but it's slow-going. 19:10:22 There are some OSA-side gate issues she's getting 19:11:02 RPM packaging - I've added a request to add jsonutils to RDO. Apparently EPEL isn't enough :(. I'll bump some of my contacts tomorrow if it doesn't get any attention soon. 19:11:20 Migration into DF - CrayZee if I recall this is on you, right? 19:11:21 Something we can help with? 19:11:32 (+1s and stuff) 19:11:43 dimak, let me see if the contacts I have come through. If not, I'll ask. 19:11:50 Cool 19:11:50 oanson: no progress with this one yet 19:11:52 I think it's mainly to approve the RDO package request 19:12:00 CrayZee, no worries. 19:12:16 I assume you'll start with a spec? 19:12:40 first thing when I get to this task :) 19:12:52 Sure. :) No rush. 19:12:55 was busy with gates fixing this week 19:13:04 Yes. Let's get to that 19:13:05 Gates - 19:13:13 CrayZee, how's our tempest? 19:13:17 Rally is working as expected 19:13:31 tempest issues - mostly fixed 19:13:40 two types of issues remain: 19:14:03 1. BGP issues - command line and containers. 19:14:23 2. One problem in the cross tenant traffic - dimak is helping me on this one 19:14:56 Yeah ,the devstack I spun up for this crashed, will respin tomorrow :( 19:14:58 All right. Let's wait with the BGP issues - we will need help from the neutron-dynamic-routing guys. 19:14:59 I am puzzled with the first and not sure we have anything to do about it 19:15:26 oanson: agree 19:15:53 It looks like it's Neutron guys. We can find out who's active and ping them. 19:16:16 I don't think there was any progress on the other gates 19:16:25 speaking of active 19:16:44 tried to get help on the opensack-oslo channel about the OSProfiler - no one answers 19:16:57 will see when their weekly is and bug them 19:17:07 That's probably a good idea 19:17:08 :P 19:17:26 I don't remember their timezone. I was sure they were mostly European... 19:17:57 Back to gates - I can probably set up a grenade job in a patch, now that I don't need project-config for that 19:18:28 s/I/we/. dimak, I don't remember how we split this one up between us? 19:18:51 Or do you want to wait for the spec to be merged, first? 19:19:05 oanson, we decided to get at it once we're happy with upgrades 19:19:18 So I'll leave it be for now 19:19:25 Let's concentrate on the spec 19:19:42 Troubleshooting - 19:20:01 skydive - no progress yet. 19:20:07 I don't have anything to update here. 19:20:23 Documentation - 19:20:39 No updates from me. But I should upload an introductory blog post to the blog. 19:21:00 Anyone wants to say something about the roadmap? 19:21:08 oanson: would you like me to write that blogpost we talked about? 19:21:14 Yes 19:21:22 k 19:21:56 #topic Bugs 19:22:06 CrayZee , secrets in compeny :) 19:22:29 :P it is a surprise 19:22:41 I've fleshed out the patches for bug 1720734 19:22:41 bug 1720734 in DragonFlow "Floating IP association to LBaaS VIP supported by Octavia does not work" [Critical,In progress] https://launchpad.net/bugs/1720734 - Assigned to Omer Anson (omer-anson) 19:23:13 The first patch in the chain: https://review.openstack.org/#/c/516101 19:23:44 It fails on the OVSDB reconnect test, even though I'm rebased above CrayZee's fix 19:24:05 please rebase it over https://review.openstack.org/527625 19:24:24 oanson, I raised a concern there about port status updates 19:24:31 you rebased it over the first patch in the patchset, use the laters 19:24:35 latest 19:24:55 dimak, I solved that today 19:25:06 CrayZee, it's rebased over master latest 19:25:09 don't see it in code :( 19:25:31 sec 19:26:10 oanson, Ok nevermind 19:26:13 dimak, line 98 and onwards in port_behind_port.py 19:26:26 :) 19:27:44 And dimak I moved the tests out of test_apps in a new patch. Since the test file should also include the regular trunk app tests 19:27:54 Since from Dragonflow point of view, it's the same application 19:28:03 oanson, Cool, will take a look 19:28:17 Cheers. Will fix the pep8 errors tomorrow 19:28:29 Hopefully I'll also find out why the fullstack tests fail :) 19:28:38 Looks good all in all, though I'll re-review in the morning. Its too late for +2's :P 19:28:47 No worries. 19:29:00 By then zuul should approve it too 19:29:10 That's all I have for bugs. Anyone else? 19:29:34 #topic Open Discussion 19:29:39 The floor is free 19:30:30 All right then. Thanks everyone for coming! 19:30:38 Cheers 19:30:42 #endmeeting