14:03:28 #startmeeting kuryr 14:03:29 Meeting started Mon Dec 11 14:03:28 2017 UTC and is due to finish in 60 minutes. The chair is irenab. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:03:30 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:03:32 The meeting name has been set to 'kuryr' 14:03:53 hellow, who is here for a meeting? 14:04:01 o/ 14:04:03 o/ 14:04:04 o/ 14:04:09 #chair dmellado 14:04:10 Current chairs: dmellado irenab 14:04:34 o/ 14:04:38 o/ 14:04:39 shall we wait for Toni? 14:04:45 o/ 14:04:48 irenab: I haven't seen him around at all today 14:04:52 I think Toni is on PTO today 14:05:05 lets start then 14:05:23 #topic kuryr-libnetwork 14:06:03 there is SRIOV patch to be reviewed #link https://review.openstack.org/#/q/project:openstack/kuryr-libnetwork 14:06:38 sorry, correct link #link https://review.openstack.org/#/c/500436/ 14:06:56 apuimedo, irenab to review 14:07:07 #topic kuryr-kubernetes 14:07:13 according to approach (https://review.openstack.org/#/c/513715/) I made patches (https://review.openstack.org/#/q/topic:bp/kuryr-kubernetes-sriov-support-refactor+(status:open+OR+status:merged)). Feel free to review 14:07:50 danil: I'll need to sign up for those. 14:08:22 dulek, irenab, others please review #link https://review.openstack.org/#/q/topic:bp/kuryr-kubernetes-sriov-support-refactor+(status:open+OR+status:merged) 14:08:42 sure, thanks 14:08:45 irenab: now you ask yourself to review too? :P 14:08:45 danil, anything specifc you want to discuss? 14:08:59 dmellado, yea :-) 14:09:09 danil: I can see that CI is failing for daemonized Cinder on the first patch. 14:09:15 nothing special. Just waiting for your comments on patches 14:09:32 danil: I'll help you to debug it and adapt. I just need to go through the patches first. 14:09:33 dulek, hm, I'll check 14:09:50 dulek, thanks, sure 14:10:40 danil, I do not have SRIOV NICs in my lab, so will review only the logic 14:10:53 irenab, ok, sure 14:11:00 danil, are we having going for a proprietary solution for the multi VIF use case ? as far as I remember we where waiting for community solution 14:11:15 danil: BTW - why don't your join #openstack-kuryr channel? :) 14:11:27 heh, true danil 14:11:28 I mean k8s community solution 14:11:30 that'd come in handy! 14:12:50 dulek, I'm there as I know. yboaron , please, give any refs for that solution. I will check 14:13:45 I think there is no clear solution yes, mostly community discussions how to keep k8s API consistent but extendable 14:13:49 yboaron , is this discussion is relevant - https://groups.google.com/forum/#!topic/kubernetes-sig-network/9YEhLBte1Z0 ? 14:14:02 danil: I don't see you in the roster. Kind of netsplit? 14:15:35 leyal, do you to give a short update on the network policy support? 14:15:53 yep 14:15:58 leyal, thanks 14:16:38 1. i will be happy to continue to discussion on the spec .. 14:17:25 2. I uploaded a WIP patch for translating k8s policy to SG (https://review.openstack.org/#/c/526916/1) 14:17:48 leyal: just for the record could you put a link to the spec around? ;) 14:18:22 #link https://review.openstack.org/#/c/519239/ 14:18:25 3. i trying to use k8s python client for serializing the network-policy object - but there is issue with the requirement .. (https://review.openstack.org/#/c/526925/) 14:18:28 thanks irenab 14:18:30 network policy support spec 14:19:15 so once requirements are updated, this should work? 14:20:16 yep, but currently support only ip_block peer type - still need to add the POD-selector and namespace-selector 14:21:01 leyal, maybe you can add a context on what would be the other patches, maybe on the launchpad board 14:21:42 irenab , will do 14:21:48 thanks! 14:21:53 o/ :) 14:22:02 janonymous, hi 14:22:34 dulek, ltomasbo , dmellado any updates? 14:22:47 irenab, not from my side 14:22:59 Not much from my side, I'm still exploring how to move VIF choice to daemon. 14:23:12 Yeah, I would love to have some reviews to gena's patch if possible, he's adding a pod to pod test and will follow up with a ports pool one 14:23:14 I'm testing it now, hope to have some code up for reviews soon. 14:23:25 https://review.openstack.org/#/c/525626/ 14:23:28 dmellado, link? 14:23:33 dmellado: It failed the gate and it's pretty difficult to debug. :( 14:23:39 #link https://review.openstack.org/#/c/525626/ 14:23:45 I’ve uploaded tests and devref to L7 routing patch . Please review : #link https://review.openstack.org/#/c/523900/ 14:23:48 dulek: I saw it, I'm trying to debug myself... 14:23:54 dmellado: And I've run it successfully on my local env, so I'm surprised. 14:23:59 it did fail on mine 14:24:03 so we have 50-50 xD 14:24:14 dmellado: I guess your env is right then. :) 14:24:23 but at least I just wanted folks to be aware of 14:24:27 and we could use some more debugging too 14:24:29 please review L7 routing patch 14:24:34 thanks! 14:24:50 10x irenab 14:26:10 are we ready to move to bugs discussion? 14:27:03 #topic bugs 14:27:44 irenab: What's about bugs? :) 14:28:05 anyone wants to discuss/raise some issue, not feature? 14:28:09 we don't have any! xD 14:28:10 lol 14:28:24 There is one I reported, not sure it is a but, but anyway https://bugs.launchpad.net/kuryr-kubernetes/+bug/1736677 14:28:25 Launchpad bug 1736677 in kuryr-kubernetes "Add native Octavia driver to support k8s services" [High,In progress] - Assigned to Irena Berezovsky (irenab) 14:29:17 Since Octavia now has its own API endpoint, it will deprecate neutron-lbaas 14:29:21 irenab: I see that we do have quite a few 14:29:31 I've reported a few, but don't have the time to work on them just now. 14:29:32 and I think it's a great idea to have this section 14:29:46 maybe we should just try to add a triage section to this meeting 14:30:02 I am not sure how urgent is to deal with it now, but something we need to be ware of 14:30:05 dmellado, agreed 14:30:31 Throwing two more reports into discussion: https://bugs.launchpad.net/kuryr-kubernetes/+bug/1734875 and https://bugs.launchpad.net/kuryr-kubernetes/+bug/1731486 14:30:32 Launchpad bug 1734875 in kuryr-kubernetes "openshift gates fail with 403 Forbidden when downloading OpenShift binaries" [Undecided,New] 14:30:33 Launchpad bug 1731486 in kuryr-kubernetes "Kuryr ignores namespaces" [High,Confirmed] 14:31:02 it ignores namespaces? oh my xD 14:31:09 dulek, I think you were also checking upgrade of devstack to k8s 1.8 14:31:16 First one should be a low hanging fruit for anyone that can read wget's manpage. 14:31:33 irenab: I'm debugging it right now, for some reason it stopped working. 14:31:57 irenab: But this is WiP patch for 1.8 support: https://review.openstack.org/#/c/525502/ 14:32:42 dulek, thanks 14:32:49 will check this 14:33:23 Let's sort out the priorities for the bugs till next meeting, to make sure what is critical to resolve 14:33:24 irenab: I think it just went through on my env, so I'll submit another version later today. 14:34:29 cool 14:35:22 #topic open discussion 14:35:51 Anyone wants to raise topic for discussion 14:35:55 ? 14:36:36 Actually… 14:36:49 What do you think about making some of the tempest gates voting? 14:37:15 I'd propose at least basic lbaasv2 and Octavia jobs. 14:37:23 Maybe lbaasv2+CNI daemon as well. 14:37:48 dulek, we can and if has issues always can brng it back to non-voting 14:37:58 dmellado, what do you think? 14:38:46 I agree on lbaasv2 and octavia 14:38:56 as dulek said, we can always put them back to nv if needed ;) 14:39:30 so +1 for making lbaasv2 and Octavia jobs voting 14:39:53 dmellado, irenab: Those two are stable. I'll file a patch and link Graphana statistics so we can check for sure that the jobs are stable. 14:39:57 +1 from me for the CNI daemon as well 14:40:08 dulek: awesome ;) 14:40:31 great! 14:40:33 irenab: Okay, I'll check it's stability in longer term, but should be similar to the aforementioned ones. 14:40:40 it's/its 14:41:03 ok, seems we are good to have extra 15 mins to attend the other meetings ;-) 14:41:52 thank you all for joining 14:42:00 #endmeeting