16:00:32 <inc0> #startmeeting kolla
16:00:36 <openstack> Meeting started Wed Oct  4 16:00:32 2017 UTC and is due to finish in 60 minutes.  The chair is inc0. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:37 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:40 <openstack> The meeting name has been set to 'kolla'
16:00:40 <inc0> #topic rollcall - w00t
16:00:52 <inc0> hello everyone
16:00:52 <britthouser> 0/
16:00:58 <duonghq> woOt
16:01:04 <duonghq> \o/
16:01:09 <spsurya__> w00t 0/
16:01:12 <krtaylor> w00t
16:02:59 <inc0> not too many of us today;)
16:03:10 <inc0> #topic announcements
16:03:30 <hrw> o/
16:03:32 <ansmith> o/
16:03:34 <inc0> 1. as you surely noticed, zuulv3 wasn't exactly smooth experience for us
16:03:47 <krtaylor> hehheh
16:04:09 <inc0> but yesterday infra rolled back zuulv2 and zuulv3 will run in pararell, acting as "3rd pary ci" for few weeks
16:04:29 <duonghq> how can we trigger v3 explicitly?
16:04:42 <hrw> so we can work on code and on gates at same time?
16:04:51 <inc0> plan is to be prepared when that happens
16:04:56 <inc0> yes
16:05:00 <inc0> https://review.openstack.org/#/c/508661/
16:06:06 <inc0> this is example of how that looks
16:06:28 <inc0> but we're not frozen till that happens'
16:06:35 <inc0> any community announcements?
16:06:58 <inc0> I see no agenda for today:)
16:07:11 <inc0> so let's just go to open discussion
16:07:15 <inc0> #topic open discussion
16:07:28 <inc0> anyone?
16:07:30 <duonghq> inc0, when the zuulv3 is triggered?
16:07:55 <inc0> hard to tell, they tried last week and run into issues
16:08:13 <inc0> as per Monty's email yesterday, 2 weeks approx
16:08:42 <duonghq> I mean atm, the v2 is the primary, and v3 is 3rd, how can we trigger v3 explicitly? or it's trigger when .zuul.yaml existed?
16:09:12 <inc0> ah, every patch will run on v3 too
16:09:14 <hrw> 508661 went to 47th version. can it get merged so it will be more visible how those tests fail?
16:09:31 <hrw> or is it tested in meantime somehow?
16:09:49 <spsurya__> inc0: just wanted to check its priority https://blueprints.launchpad.net/kolla-kubernetes/+spec/devenv-taketwo    as we already have two alternate dev env for kolla-k8s
16:09:55 <duonghq> it's still fail on both centos-source test on zuulv3, sadly
16:10:03 <spsurya__> so should i continue with this ?
16:10:07 <inc0> well, that's my fault;)
16:10:15 <inc0> me and Jeffrey hack around it
16:11:01 <duonghq> ya, I think we should wait for zuulv3 is officially before merge?
16:11:17 <inc0> no, because when it does, we won't have any gates
16:11:29 <inc0> we should work on this patch over next weeks
16:11:31 <hrw> we need v2 and v3 running before v2 gets killed
16:11:38 <inc0> so when the switchoff happens, we'll be ready
16:11:40 <hrw> right?
16:11:42 <inc0> yes
16:11:46 <duonghq> roger
16:12:11 <inc0> spsurya__: devenv for kolla-k8s?
16:12:16 <spsurya__> yes
16:12:27 <duonghq> https://review.openstack.org/#/c/425446/ -> may somebody help me test Keystone upgrade logic?
16:12:39 <inc0> well, for sure quickstart needs revisit
16:12:45 <spsurya__> ansible one already progress
16:12:49 <inc0> I know for a fact it's stale and wrong
16:12:50 <inc0> ahh
16:12:53 <inc0> that's different
16:13:07 <inc0> ansible - using kolla-ansible to develop openstack services like heat
16:13:15 <inc0> k8s - developing kolla-k8s itself
16:14:14 <inc0> so back to gates - yes we need to have v3 working before v2 gets killed
16:14:27 <spsurya__> inc0:  i will keep it on low priority
16:14:40 <spsurya__> i will discuss with sdake too
16:14:47 <kfox1111> kolla-kubernetes's gates work except for multinode (likely iptables changes) and an ironic gate.
16:14:48 <spsurya__> this was his idea
16:14:58 <inc0> kfox1111 will be better person to talk to;)
16:15:16 <inc0> if you want to do it Surya, you're welocme to
16:15:23 <spsurya__> kfox1111: morning
16:15:24 <duonghq> inc0,  atm, I think we should just migrate from v2 to v3, with exactly the same jobs in v2?
16:15:27 <kfox1111> morning. :)
16:15:35 <inc0> duonghq: yeah
16:15:46 <spsurya__> kfox1111: regarding this https://blueprints.launchpad.net/kolla-kubernetes/+spec/devenv-taketwo
16:16:07 <kfox1111> spsurya__: you interested in implementing that?
16:16:09 <inc0> but they're not identical
16:16:21 <spsurya__> kfox1111: yes I am
16:16:29 <kfox1111> cool.
16:16:51 <spsurya__> kfox1111: seems like i have to with minikube
16:16:54 <kfox1111> I'm still recommending minikube for that use case. as a dev can very easily burn down/redeploy as needed.
16:17:07 <kfox1111> yeah. I think thats a better aproach.
16:17:19 <spsurya__> kfox1111: roger, thanks
16:17:23 <kfox1111> oh.
16:17:31 * hrw off
16:17:33 <kfox1111> I played around with rook in minikube a little while ago.
16:17:40 <kfox1111> it worked well.
16:17:52 <kfox1111> not production ready, but should be just fine for dev stuff.
16:18:03 <kfox1111> so might be an alternate way to get ceph in the minikube.
16:18:28 <spsurya__> kfox1111: yes
16:18:43 <inc0> deploy ceph with ansible and just connect to it?;)
16:19:08 <spsurya__> inc0: kfox1111 https://blueprints.launchpad.net/kolla-kubernetes/+spec/multinode-deployment-guide
16:19:26 <kfox1111> inc0: probaby woudlnt work in minikube.
16:19:34 <kfox1111> I dont think python'sinstalled in the vm.
16:20:12 <kfox1111> spsurya__: looks good to me. you wana work on that too?
16:20:47 <spsurya__> kfox1111: AIO deployment guide will be base for it
16:20:47 <inc0> ok guys, I think we can end meeting and move to regular channel
16:21:08 <inc0> thank you all, have 40min back of your lifes
16:21:10 <kfox1111> k.
16:21:14 <inc0> #endmeeting kolla