14:00:30 #startmeeting kuryr 14:00:31 Meeting started Mon Apr 23 14:00:30 2018 UTC and is due to finish in 60 minutes. The chair is dmellado. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:32 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:35 The meeting name has been set to 'kuryr' 14:00:41 Hi kuryrs, who's here today for the meeting? 14:02:03 o/ 14:02:15 oh, I'm not alone! xD 14:02:17 hi dulek 14:02:21 Hey. 14:02:22 o/ 14:02:24 Hi guys 14:02:25 #chair dulek 14:02:25 Current chairs: dmellado dulek 14:02:28 #chair dulek yboaron 14:02:29 Current chairs: dmellado dulek yboaron 14:03:16 563137 is finishing check queue and will be getting into the merge one. It's not a lot of tests there on stable/queens, so fingers crossed. ;) 14:04:09 #topic kuryr-kubernetes 14:04:13 soooo, RELEASE 14:04:55 I'll prepare a release commit once this last fix merges. 14:05:10 just for overall information we were planning to do a release on Friday 14:05:19 Then it's dmellado +1 needed on it and bugging the release folks. 14:05:21 but we've been struggling with some issues, hopefully sorted out by now 14:05:27 yeah, I'll take care of that ;) 14:05:39 so, overall issues, devstack wasn't applying any caps on the tempest plugin reqs 14:05:46 so they went haywire and broke all the things 14:05:54 yboaron: that's what you saw on those patches 14:06:06 dmellado, yep! 14:06:15 we fixed that and then backported 14:06:17 https://review.openstack.org/#/q/dc5d88bc0b2233ee8490d31ce67d6a8f6e503b23 14:06:45 also, centos amphora had some issues that made it not work properly 14:06:52 dmellado, thanks for the update! 14:06:53 first one was due to DIB broke due to pip10 14:07:11 the image builder was broken so no amphora image so boom 14:07:25 so we decided to go for their qcow2 published image 14:07:49 which had some issues on connectivity on L2 14:07:59 after switching to ubuntu, things seems to be working now 14:08:16 those amphora issues made all service tests fail 14:08:31 so we had quite the interesting friday 14:08:40 dmellado,How the K8S python client version was related to those issues ? 14:08:42 thanks dulek for all your work into this too 14:08:46 yboaron: that's another thing 14:08:50 and comes from the tempest plugin 14:08:56 as it had no UC cap 14:08:58 yboaron: Issues always come in pairs. :P 14:08:59 dmellado, you had a very busy weekend :-) 14:09:02 it installed a newer version 14:09:19 from the python k8s client which changed how the service was created 14:09:24 it introduced a new parameter IIRC 14:09:33 and thus the test broke 14:09:47 I guess that's the overall story 14:09:49 o/ (forgot about this...) 14:09:51 BUT! 14:10:01 today we can finally get to release 14:10:13 and as a side and funny effect just go and -1 ltomasbo patches 14:10:25 because of him being late, but that's another story 14:10:29 xD 14:10:42 ok, so that's all for me on the release side, how about you folks? 14:10:59 #chair ltomasbo 14:10:59 Current chairs: dmellado dulek ltomasbo yboaron 14:11:09 I've started looking at fun thing. :) 14:11:29 Basically I wanted to try https://kubernetes.io/blog/2016/01/simple-leader-election-with-kubernetes for leader election of kuryr-controller. 14:11:56 I'll take a look dulek 14:12:07 for the channel could you spare a quick summary on it? 14:12:09 I'm a little bit worried how this works. Basically there's a side-car container that does the leader-election. 14:12:42 And I don't really like depending on a container coded in Kubernetes repo and hosted by Google on GCR. 14:13:12 dulek: do you have a spec on this? 14:13:14 On the other hand rewriting that in Python (NIH) doesn't make too much sense either. 14:13:28 dmellado: Not yet, just wanted to see how this works. 14:13:51 dmellado: A POC should be pretty easy to show, so I hope to have something soon. 14:13:53 dulek: ack, I'll take a look too and might be asking you for details so I can properly reply after it 14:13:57 dulek: even better ;) 14:14:05 Yup. :) 14:14:15 Okay, that's it from my side. 14:14:54 garyloug: yboaron any updated from the DPDK side? 14:16:10 dmellado, I've just been making some changes based on the feedback I'm getting 14:16:21 not much update 14:17:19 yboaron: thanks for handling the reviews so far 14:17:33 garyloug: let us know if you're missing any reviews and I'll push for those 14:17:35 dmellado, your'e welcome! 14:17:50 garyloug, Q: how the the DPDK interface being attached to the relevant Pod ? 14:18:09 garyloug, I mean after we've excluded it from the network stack 14:20:05 yboaron, this info is passed within the annotation 14:20:50 we use downward API mechanism in K8s to get the DPDK interface 14:21:27 I will add this in more detail on the patch itself 14:21:42 garyloug, thanks! 14:21:43 sorry i never got around to replying to you 14:21:52 garyloug, NP! 14:21:59 garyloug: awesome, if you hadn't already please add some docs within the patch as well 14:22:22 dmellado, yes, it's on my todo list 14:23:02 garyloug: thanks! 14:23:15 NP :) 14:23:16 ltomasbo: anything from your side? 14:23:36 working on the network per namespace patch sets 14:23:52 still one to finish the complete set (handling proper ports pool deletion) 14:23:58 and then I need to add unittests 14:24:11 so, feel free to take a look and provide reviews! 14:24:27 other than that, trying to have a better coverage for octavia and ovs-firewall at the gates 14:24:45 as we were not testing a few things and that is why we had some issues the last fridays... 14:24:53 ltomasbo: I've seen your patches on the gates, if I'm right there's an actual bug on zuul v3 parsing 14:24:57 that's all on my side 14:25:05 so, more infra side 14:25:14 for the ovs-firewall, right? 14:25:21 I'm waiting on this patch to get finally merged to enable multi node testing 14:25:25 ltomasbo: yeah 14:25:32 yep 14:25:49 https://review.openstack.org/#/c/545724/ 14:25:59 this is the one we need merged to enable or gates, but those are done 14:26:55 besides this, I also wanted to tackle network policy support 14:26:57 so 14:26:59 irenab: around? 14:27:47 all right, not yet 14:27:59 so, all the patches on network policy support have been there for a while 14:28:13 and I think I'll be taking them over, rebase them and give them a try 14:28:25 so we can get to merge them and create a new gate/tests for this 14:28:35 I'll be setting a VC this week to discuss this 14:28:38 https://groups.google.com/forum/#!topic/kubernetes-sig-network/A8oyMrWpJ8Q 14:28:47 please do take a look at ^^ prior to the meeting 14:28:56 in case you're interested in attending 14:31:20 besides this, I'll be creating several slides for the presentations that we do have for the Vancouver summit 14:31:35 and I might be asking for collaborations from you , as you're just scared of flying there 14:32:02 #topic open discussion 14:32:16 Anything else around that you'd like to speak about, folks? 14:33:00 Maybe it's worth speaking about your PTG email? 14:33:12 At least making people aware of. :) 14:33:46 yeah, that came the final thing 14:33:48 xD 14:37:54 sorry folks, did you get to see my email about the next PTG? 14:40:35 so please reply before May 2nd 14:40:37 https://goo.gl/forms/HfiNnEF2CwuMva6n1 14:40:53 dulek: ltomasbo yboaron irenab garyloug 14:40:57 and whoever ;) 14:41:02 I've answered already. :) 14:41:25 thanks dulek ;) 14:41:33 snapiri: ^^ :P 14:44:03 #endmeeting kuryr