14:02:21 <dulek> #startmeeting Kuryr
14:02:22 <openstack> Meeting started Mon Apr  9 14:02:21 2018 UTC and is due to finish in 60 minutes.  The chair is dulek. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:02:26 <openstack> The meeting name has been set to 'kuryr'
14:02:36 <dulek> Hi people, who's here for the meeting?
14:02:56 <dulek> irenab, dmellado, ltomasbo, celebdor?
14:03:29 <ltomasbo> I'm here (half of me as I'm in another call too)
14:03:33 <irenab> hi, I am here, but have to leave in 10 mins
14:03:42 <openstack> irenab: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
14:03:51 <ltomasbo> xD
14:03:54 <dulek> :)
14:03:55 <ltomasbo> nested kuryr meeting
14:03:59 <irenab> :-)
14:04:15 <dulek> Let's wait a bit for celebdor.
14:04:25 <garyloug> o/
14:05:23 <dulek> dmellado is not going to join today, but he left us with some discussion points.
14:05:35 <celebdor> dulek: Here I am
14:05:36 <celebdor> sorry for the delay
14:05:45 <dulek> So as we have garyloug here we can start with DPDK patch.
14:05:46 <celebdor> my irc bridge was misbehaving
14:05:52 <dulek> #link https://review.openstack.org/#/c/559363
14:05:55 <irenab> any chance we can move the meeting 1 hour earlier?
14:06:38 <dulek> I think RH folks have conflict an hour earlier. At least I have.
14:06:52 <dulek> But I'm fine with looking for a new spot if this one doesn't work.
14:07:21 <irenab> never kind, will try to figure out how to rearrange
14:07:39 <dulek> Okay.
14:08:02 <dulek> So garyloug, do you have anything to say about the DPDK patch?
14:08:13 <garyloug> Hi dulek
14:08:22 <dulek> Questions? ETA? Anything we should focus on when reviewing?
14:08:58 <garyloug> Yea, so patch is work in progress, I have a few TODOs in the code. I also need to do unit tests and tempest stuff.
14:09:07 <celebdor> dulek: did you start a topic?
14:09:24 <dulek> celebdor: Hm, not really.
14:09:29 <apuimedo> ok
14:09:32 <dulek> #topic Kuryr-Kubernetes - DPDK patch
14:09:49 <apuimedo> garyloug: what's the status?
14:09:51 <dulek> #chair irenab apuimedo
14:09:51 <openstack> Current chairs: apuimedo dulek irenab
14:10:47 <garyloug> apuimedo, not sure if you cought me comment above, it's work in progress
14:10:50 <apuimedo> garyloug: I saw while on holidays that you put w-1
14:11:05 <apuimedo> no, I sdidn't see any comment
14:11:27 <garyloug> yes, I have some TODOs still in the code. I also have to do unit tests and tempest
14:11:30 <apuimedo> would you like us already to take a look or you prefer it after the WIP
14:11:35 <apuimedo> ok
14:11:47 <garyloug> yes, please take a look :)
14:12:06 <dulek> All right, thanks for update!
14:12:20 <irenab> garyloug, is it based on the multi vif patches?
14:12:54 <apuimedo> will do, thanks
14:13:13 <garyloug> irenab, I've not tried it with multi vif patches, but this is the end goal I guess
14:13:25 <garyloug> what is the status of these patches?
14:14:04 <irenab> I didn't see any progress lately, but was out for a while, so maybe missed something
14:14:52 <garyloug> yea, I've been out also :) So if there's any update guys please let me know.
14:15:13 * irenab have to leave, will sync up on logs later
14:15:38 <apuimedo> irenab: very well
14:15:50 <apuimedo> #topic kuryr-kubernetes cni exec
14:16:15 <dulek> apuimedo: Any findings on the VM?
14:16:24 <apuimedo> #info we have some issues for the host namespace to be able to reach the loadbalanced API server. We are currently debugging on an infra VM
14:16:29 <apuimedo> dulek: first I thought it was SGs
14:16:47 <dulek> apuimedo: I was thinking about that as well, but DevStack's default SGs are permissive.
14:17:00 <apuimedo> because there was no SG allowing comm to the LB from the pod network
14:17:04 <dulek> apuimedo: And TCP connection is established.
14:17:22 <apuimedo> (you fixed this for Kuryr created LBs but it is still not handled in devstack)
14:17:35 <apuimedo> somehow it does not seem to be enough
14:17:40 <apuimedo> so I'm still digging into it
14:17:53 <dulek> apuimedo: Ah. Still LBaaS port will get default SG, so it shouldn't be the case.
14:18:02 <apuimedo> dulek: actually
14:18:05 <dulek> apuimedo: Because DevStack's default is allow all.
14:18:15 <apuimedo> the devstack created one is in a different 'default' sg
14:18:17 <apuimedo> :-D
14:18:28 <apuimedo> there's like 4 'default' sg
14:18:42 <dulek> apuimedo: Yeah, I think it's for each project.
14:18:56 <apuimedo> and the devstack LB one is different from the 'default' we have in pods (and kubelet)
14:19:22 <dulek> Interesting.
14:19:35 <apuimedo> indeed
14:19:40 <apuimedo> I expect to find it today though
14:19:54 <apuimedo> ltomasbo: do you have some topic?
14:20:12 <dulek> I'll be super grateful, it took way too much time.
14:20:33 <dulek> And if we're talking about gates, here's dmellado update on Multinode:
14:20:33 <dulek> I'm oddly stuck when assigning the host-vars for the controller, finding out that it won't create the .kube/config file. I'll do some experiments today but might need to duplicate the tempest service install in both vars and host-vars. As it happened to Michal, whenever I thought this would be an easy fix it's taking longer than I expected.
14:21:21 <ltomasbo> not really, just pushed some backports and the documentation about the kuryr-kubernetes version vs kubernetes one
14:21:24 <ltomasbo> apuimedo, ^^
14:21:48 <apuimedo> thanks ltomasbo
14:22:03 <apuimedo> I'll try to look at the .kube/config dmellado thing tomorrow
14:22:46 <dulek> #topic Kuryr-Kubernetes - network policies
14:23:14 <dulek> dmellado said that Eyal is leaving Huawei and dmellado will take over the patches that are left.
14:23:19 <dulek> #link https://review.openstack.org/#/q/status:open+project:openstack/kuryr-kubernetes+branch:master+topic:bp/k8s-network-policies
14:24:18 <dulek> I don't really have any comments beside that it would be nice to get some reviews there (mine included).
14:24:31 <apuimedo> dulek: dmellado will get really busy
14:24:33 <apuimedo> :-)
14:24:42 <apuimedo> dulek: agreed on the review side
14:24:50 <apuimedo> I should take a look as well
14:24:56 <apuimedo> as soon as I have coding time though
14:25:02 <apuimedo> I'll be fixing neutron speed
14:25:06 <apuimedo> for bulk
14:26:13 <apuimedo> do we have any other topic?
14:26:18 <dulek> #topic Kuryr-Kubernetes testing
14:26:26 <apuimedo> oh, ywah
14:26:38 <dulek> dmellado also wanted to raise this patch https://review.openstack.org/#/c/558482
14:26:38 <apuimedo> what's with the ovn gates ltomasbo ?
14:27:03 <apuimedo> cool
14:27:09 <ltomasbo> I didn't have time to look at it, but the problem was with security groups between VM tempest network and pod tempest network
14:27:19 <ltomasbo> only allowing communication one way, not the other
14:27:49 <ltomasbo> either tomorrow or on thursday I will try to take a look with dmellado to that one
14:29:54 <dulek> Okay, anything else to discuss?
14:30:03 <dulek> #topic open discussion
14:30:36 <apuimedo> I got nothing
14:31:22 <dulek> So I guess we can cut short today. :) Thanks everyone!
14:31:28 <dulek> #endmeeting