08:00:12 <oanson> #startmeeting Dragonflow
08:00:13 <openstack> Meeting started Mon Dec 11 08:00:12 2017 UTC and is due to finish in 60 minutes.  The chair is oanson. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:00:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:00:16 <openstack> The meeting name has been set to 'dragonflow'
08:00:19 <snapiri> Hi
08:00:24 <oanson> Yo
08:00:26 <dimak> Good morning
08:00:38 <lihi> Hi
08:00:46 <oanson> All right. Let's start
08:00:49 <oanson> #topic Roadmap
08:01:13 <oanson> LBaaS - no update. Spec has been merged a couple of weeks back
08:01:17 <oanson> DNS - lihi ?
08:02:27 <oanson> Until lihi returns - RPM packaging - I god jsonmodels into fedora. Now trying to get it into centos. If anyone can test and vote for https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2017-77f192449a that would be great!
08:02:32 <leyal> \o
08:02:39 <oanson> leyal, hi
08:03:12 <lihi> oanson, our basic integration with designate works out of the box
08:03:25 <lihi> I'm working on a spec to discuss our next options
08:03:43 <oanson> Cool. You want to bring it up for discussion here? Or you prefer to wait for the spec?
08:04:16 <lihi> I think it needs context. Therefore the spec :)
08:04:47 <oanson> Sure
08:05:31 <oanson> Next up: Deployment.
08:05:38 <oanson> lihi, your up again. OSA ?
08:06:08 <lihi> The test are still failing, but I'm working on it. It takes few hours for them to run, but I hope to make it work this week
08:06:49 <oanson> Did you try a full deployment as well?
08:07:22 <lihi> Yes
08:07:35 <lihi> This is what I'm testing
08:07:50 <oanson> Ah sure. I thought you meant the tox environment
08:08:26 <oanson> All right.
08:08:37 <oanson> I don't think there is any progress with the gates.
08:09:30 <oanson> Except maybe with the Kuryr integration gate. I understand irenab is in contact with dmellado. I guess they'll update next time they're in. (As we said before, we plan to just take their tests and run them under Dragonflow)
08:09:40 <oanson> Troubleshooting
08:09:57 <snapiri> oanson: I believe we know why the gates fail, there is not much to do about it
08:10:01 <dmellado> oanson: o/
08:10:07 <oanson> dmellado, hi
08:10:18 <dmellado> just saw the notice, did I miss something? ;)
08:10:34 <oanson> dmellado, no. Just mentioning you regarding the kuryr tests
08:10:44 <dmellado> ack, we're adding some more tests there and I'm working on the multinode gates
08:10:48 <oanson> We're waiting to see how well it works under Dragonflow :)
08:10:48 <dmellado> including dragonflow
08:10:57 <oanson> Cool!
08:10:58 <dmellado> I expect to have something hopefully this week ;)
08:11:07 <oanson> Very cool!
08:11:27 <oanson> Thanks! :)
08:11:41 <oanson> snapiri, you mean the tempest tests?
08:11:47 <snapiri> and rally
08:12:21 <oanson> Let's take them one at a time
08:12:26 <oanson> And start with tempest
08:12:28 <oanson> :)
08:12:31 <snapiri> :)
08:13:04 <oanson> What did you see with the tempest tests? Why is it something we can't solve?
08:13:14 <snapiri> so tempest fails due to BGP issue
08:13:31 <dimak> We can turn the tests off if they just introduce noise
08:13:44 <snapiri> it fails to parse some cli output (do not remember exactly what)
08:14:27 <snapiri> Can't recall at the moment, but it is in some internal package...
08:14:28 <oanson> This is a problem we may have to solve with the neutron-dynamic-routing people
08:14:50 <snapiri> I believe it is in ryu
08:15:01 <oanson> All right. Let's put BGP on the to-do list. Anything else with the tempest tests? There are other tests that are unstable. Did you map them?
08:15:18 <snapiri> this is what fails us when it fails.
08:15:28 <snapiri> nothing else I see,
08:15:43 <snapiri> on the rally it is something totally different
08:16:20 <oanson> Sure. What's the issue with rally?
08:17:10 <snapiri> the list-ports takes a bit over the time that is set as maximum
08:17:43 <snapiri> we create much more ports than the regular test, and still set the maximum time to the same value
08:18:15 <snapiri> This is a piece of code that does not go through Dragonflow
08:18:29 <oanson> Then we can increase the timeout. Does list-ports even go through Dragonflow code?
08:18:34 <oanson> Beat me to it :)
08:18:38 <snapiri> depending on the VM and host load, it may fail or succeed
08:18:42 <oanson> Sure. Then let's increase the timeout there and move on
08:19:04 <oanson> Or even remove the entire list-port test. It isn't relevant to us.
08:19:19 <oanson> I don't want to test Neutron on our gates. Neutron have their own gates for that :)
08:19:21 <snapiri> it is a part of a bigger test
08:19:33 <oanson> I See
08:19:37 <oanson> see*
08:20:19 <snapiri> let's take it offline and handle it today...
08:20:22 <oanson> Sure.
08:20:36 <oanson> As a first step - let's increase the timeout and gauge the stability
08:20:44 <oanson> Troubleshooting
08:20:57 <oanson> snapiri, this entire section is yours :)
08:21:13 <oanson> So choose what you want to start with, and go ahead
08:21:24 <snapiri> df-model?
08:22:09 <snapiri> It has been improved and now can generate output of our schema to rst format as well (can be used in documentation).
08:22:47 <oanson> Cool
08:22:55 <snapiri> Also the vagrant install was fixed, so now we can set-up a controller+2*compute nodes pretty easily
08:23:11 <oanson> re: df-model - it can also do UML, and jsonschema (for OpenAPI), right?
08:23:31 <snapiri> and plaintext :)
08:24:00 <oanson> Cool. And cool re the vagrant thing
08:24:12 <snapiri> vagrant - it takes much more time than before, though, as the building of the VMs is serial and not in parallel
08:24:38 <oanson> I see
08:25:05 <oanson> Did you manage to get a hold of the OSProfiler guys?
08:25:37 <snapiri> My bad - just tried today, hope to have answers by EOD
08:26:18 <oanson> I personally think the patch is good enough to be merged.
08:26:26 <oanson> Are there any objections to review it?
08:26:56 <dimak> I'll go over it today
08:27:11 <oanson> Remember that in Open Source, the ship never leaves the marina. We can always come back and fix the code if there are major bugs we miss during review
08:27:23 <oanson> Or even minor ones.
08:27:25 <oanson> Any bugs, really
08:27:43 <dimak> If it never leaves the marina, we never know if its seaworthy :P
08:27:43 <oanson> Documentation
08:27:53 <snapiri> I prefer it being reviewed, though :)
08:28:25 <oanson> dimak, please don't punch holes in my metaphor. It might sink :)
08:28:29 <oanson> Documentation
08:28:31 <dimak> ;)
08:28:36 <oanson> Is there anything to report for documentation?
08:28:44 <dimak> Don't have updates
08:28:53 <dimak> I promised to reshape the TOC
08:28:53 <oanson> Anything for roadmap before we continue?
08:28:56 <dimak> Didn't get to it yet
08:29:16 <oanson> dimak, sure. Thanks!
08:29:39 <oanson> Anything else for roadmap before we continue?
08:30:05 <oanson> #topic Bugs
08:30:26 <oanson> I'm working on this one: https://bugs.launchpad.net/dragonflow/+bug/1720734
08:30:27 <openstack> Launchpad bug 1720734 in DragonFlow "Floating IP association to LBaaS VIP supported by Octavia does not work" [Critical,New] - Assigned to Omer Anson (omer-anson)
08:30:36 <oanson> I uploaded all the patches needed to solve it
08:30:59 <oanson> This is the top of the chain: https://review.openstack.org/#/c/527008
08:31:10 <oanson> And this is the topic: https://review.openstack.org/#/q/topic:bug/1720734+(status:open+OR+status:merged)
08:31:20 <oanson> It even includes full stack tests now :)
08:31:39 <dimak> Cool, I'll review today as well
08:31:49 <oanson> Great. Thanks!
08:31:52 <oanson> Anyone else working on interesting bugs they want to discuss?
08:32:34 <oanson> All right.
08:32:37 <oanson> #topic Open Discussion
08:32:46 <oanson> The floor is for the taking
08:34:14 <oanson> All right then.
08:34:20 <oanson> Thanks everyone for coming
08:34:35 <oanson> #endmeeting