20:06:44 <sdake> #startmeeting kolla 20:06:45 <openstack> Meeting started Mon Oct 20 20:06:44 2014 UTC and is due to finish in 60 minutes. The chair is sdake. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:06:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:06:48 <openstack> The meeting name has been set to 'kolla' 20:06:50 <sdake> #topic rollcall 20:06:54 <larsks> Still here. 20:06:57 <sdake> \o/ hi folks take 2 :) 20:06:58 <jlabocki> me too 20:07:01 <jpeeler> hello again 20:07:01 <radez> hey ya'll 20:07:01 <daneyon> here 20:07:05 <bth-> heynow 20:07:07 <rhallisey> hey 20:07:17 <rbowen> ¯\_(ツ)_/¯ 20:07:38 <sdake> #topic agenda 20:07:46 <sdake> #link https://wiki.openstack.org/wiki/Meetings/Kolla 20:07:55 <sdake> anyone have anything to add? 20:08:03 <sdake> also if you do have stuff to add feel free to edit that wiki page 20:08:58 <sdake> #topic milestone 1 blueprint review 20:09:07 <sdake> #link https://blueprints.launchpad.net/kolla/milestone-1 20:09:28 <sdake> almost all green 20:09:34 <sdake> nvoa has some outstanding reviews 20:10:16 <sdake> https://review.openstack.org/#/q/project:+stackforge/kolla,n,z 20:10:37 <sdake> so we need to do a quick review of heat 20:10:45 <sdake> I've been reviewing along the way and it looks pretty solid 20:10:55 <sdake> larsks can you take a look at the review for heat? 20:11:03 <sdake> https://review.openstack.org/128994 20:11:08 <larsks> Sure; I will do that this afternoon. 20:11:11 <sdake> https://review.openstack.org/129639 20:11:16 <larsks> (errrr.../evening) 20:11:24 <sdake> radez as well if possible? 20:11:35 <radez> sure thing 20:11:51 <sdake> cool 20:12:01 <sdake> thgat just leaves https://review.openstack.org/129639 20:12:03 <sdake> and we are done 20:12:32 <sdake> jpeeler be watching for when your patch lands and change the state in launchpad 20:12:41 <jpeeler> yep, on it 20:12:52 <sdake> larsks can you have a review of https://review.openstack.org/129639 20:12:57 <sdake> and daneyon and jpeeler as well ? :) 20:13:00 <sdake> then I think we are set 20:13:12 <jpeeler> well i submitted that one too 20:13:29 <sdake> there are no filed bugs, so I'll skip the bug review 20:13:33 <daneyon> i'll review both after the meeting. 20:13:35 <sdake> anyone have any comments on the blueprint review? 20:13:39 <sdake> thanks daneyon 20:13:53 <sdake> #topic milestone #1 announcement 20:14:01 <sdake> #link https://etherpad.openstack.org/p/kolla-1-ann 20:14:24 <sdake> ok what would be ideal is if we had 2-3 folks run through the instructions and make sure they work 20:14:27 <sdake> any volunteers? 20:14:37 <sdake> needs to be done in next 2 hours 20:14:53 <jlabocki> I'll do it 20:14:54 <larsks> Not it. 20:15:03 <jlabocki> as long as someone is available on the tech support line 20:15:04 <radez> I won't be able to, sry 20:15:04 <jlabocki> :) 20:15:12 <jlabocki> link me to the instructions 20:15:18 <bth-> i can do it as well 20:15:22 <sdake> they are in the last link 20:15:29 <rhallisey> sdake, I can do it too 20:15:40 <sdake> cool so jlabocki, bth, rhallisey 20:15:47 <daneyon> etherpad announcement looks sweet. 20:15:49 <sdake> I'll also give it a spin 20:15:55 <larsks> I think it is important to emphasize in any announcement the early state of this work, and that we do not, in fact, expect it to produce a completely functional environment. 20:15:58 <sdake> daneyon group effort :) 20:16:08 <larsks> Otherwise we will be inundated with "it didn't work" messages. 20:16:15 <sdake> larsks agree - if you want to add the necessary disclaimer that would be good 20:16:22 <sdake> but people will often only take a look at software once 20:16:32 <sdake> so we want them to look and *understand* what it offers long term for them 20:17:08 <sdake> any other thoughts on the announcement and final testing? 20:17:25 <sdake> bth, jlabocki, rhallisey if you can ping me via irc when your done Id appreciate it 20:17:31 <rhallisey> sure 20:17:36 <jlabocki> will do 20:17:37 <bth-> ack 20:18:03 <sdake> #topic Milestone #2 blueprint review 20:18:12 <sdake> #link https://blueprints.launchpad.net/kolla/milestone-2 20:18:12 <sdake> * Open Discussion 20:18:16 <sdake> oops :) 20:18:19 <sdake> so milestone 2 20:18:22 <sdake> there is a bunch in there 20:18:28 <sdake> we will porbably have to push off stuff to milestone 3 20:18:52 <sdake> lets do a live prioritization discussion of milestone #2 20:18:58 <sdake> if anything is wrong priority please let me know 20:19:03 <daneyon> ya, DS prep is going to slow some things down for me. 20:19:06 <sdake> I will use priority to cut stuff into milestone #3 20:19:16 <sdake> DS prep? 20:19:22 <daneyon> Paris 20:19:24 <sdake> oh you mean oDS 20:19:51 <sdake> anything in essential that hsould be in high? 20:19:58 <larsks> I would push off barbican and zaraq right away, since those are still in incubation (so...nice to have, but not essential). I would also push out k8s-reset-librarb and kolla-cli probably. I would make horizion higher priority. 20:20:05 <larsks> s/zaraq/zaqar/ 20:20:36 <larsks> I would prioritize config-outside-container, because that will make it much easier for us to scale things out. 20:20:45 <rhallisey> sdake, how about ironic? 20:20:53 <larsks> We need blueprints for figure-out-the-heck-we-are-doing-with-networking. 20:21:03 <radez> I agree I think the core projects shoudl be higher prio than the incubation ones 20:21:10 <jlabocki> larsks: +1 on networking 20:21:19 <rhallisey> agreed larsks 20:21:32 <sdake> kolla rest lib and kolla cli are manage outsid ethe consider features larsks 20:22:03 <jlabocki> I would put less emphasis on cli and rest and more on making the core services work 20:22:12 <larsks> sdake: Those things do not related to the config-outside-container blueprint. 20:22:33 <sdake> oh i se 20:22:35 <larsks> jlabocki: Right, yes. I think those should be milestone-3 at least. 20:22:36 <sdake> I missed that 20:23:05 <sdake> ok any other thoughts? 20:23:16 <sdake> I'll take all this feedback and sort into a m2/m3 list 20:23:21 <daneyon> Thoughts on the following for neutron: 1. Disable L3/DHCP name spaces and allow_overlapping_ips= false. Use ML2+LinuxBridge/L2 Pop mech drivers and VXLAN tenant networks? 20:23:25 <sdake> (wednesday likely) 20:24:00 <larsks> sdake: We should probably have a blueprint for deployment orchestration: we have lots of environment variables, potential storage configuration choices, etc. How is this all set before starting up pods? 20:24:11 <sdake> larsks that is what the cli is for 20:24:18 <sdake> but I'll maek it more clear with an additional blueprint 20:24:38 <jpeeler> when is gating going to become voting? and what about testing in general? 20:24:39 <sdake> that is the sole purpose of the cli/restapi blueprints 20:24:43 <daneyon> btw-: found #link https://github.com/dave-tucker/docker-ovs for ovs in userspace. 20:25:05 <sdake> let me switch to open discussion for that jpeeler 20:25:14 <sdake> still talking about priorities on blueprints for milestone 2 / 3 20:25:46 <daneyon> My though is linuxbridges instead of ovs until we can test ovs in userspace. 20:25:51 <sdake> #topic open discussion 20:26:12 <larsks> sdake: have you had the chance to fix the bashate gate yet? 20:26:17 <larsks> I haven't checked recently. 20:26:17 <sdake> nada 20:26:25 <sdake> been working like a maniac for 2 weeks 20:26:30 <sdake> i'll get to it afer this release 20:26:34 <sdake> probably when folks are at summit 20:27:17 <sdake> regarding jpeelers gate question 20:27:22 <sdake> we really need to get a gate in place soon 20:27:25 <sdake> that does actual testing 20:27:33 <sdake> I'm not s ure how best to do that 20:27:39 <sdake> I guess we could use t empest 20:28:15 <jlabocki> votes for using openshift origin :) 20:28:26 <sdake> daneyon I don't have pref on the network tech used as long as it works like a champ :) 20:28:50 <sdake> any other open topics? 20:29:14 <larsks> kubernetes 0.4 is in fedora updates-testing, coming soon to a deployment near you. 20:29:18 <sdake> #topic release steps 20:29:20 <sdake> ok so next steps are 20:29:25 <larsks> Hold on there, cowboy. 20:29:31 <sdake> #undo 20:29:32 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x33b9dd0> 20:29:33 <daneyon> larsks: do you have any thoughts/concerns with the ideas i threw our re neutron? 20:30:13 <larsks> daneyon: I think it sounds reasonable? I would prefer kernel OVS, and just require OVS on the host, but if it can work in userspace I guess it's worth a shot. 20:30:29 <larsks> But ultimately, let's get something that works, and we can always re-jigger things as necessary. 20:30:55 <daneyon> larsks: got it. thx. 20:31:02 <larsks> sdake: Actually, is this the right time for talking about kube 0.4? Or do you want me to hold off until later in the meeting... 20:31:30 <sdake> now is good 20:31:41 <larsks> Okay. So, 0.4 is coming. It will totally break existing deployments. 20:31:53 <rhallisey> :( 20:32:00 <larsks> The "kubecfg" command is gone, replaced by "kubectl". There are some new requirements for our YAML files. 20:32:20 <larsks> I am going to looking at fixing the heat templates tomorrow. 20:32:50 <larsks> You can test it now with "yum --enablerepo=updates-testing upgrade", but you will need to deal with the new config stuff in /etc/kubernetes. 20:33:00 <larsks> ...and who knows what else. Still looking into it. 20:33:08 <larsks> I will send email to the ML when I have things working. 20:33:20 * larsks is all done with k8s 0.4 topics. 20:33:35 <sdake> anything else folks? 20:33:44 <sdake> I just wanted to cover real quick our tactical steps for the release 20:33:50 <sdake> 1. merge existing patches 20:34:05 <sdake> 2. 3 testers identified to manually test release instructions 20:34:13 <sdake> 3. tag and release 20:34:22 <sdake> 4. announce to ml tmorrow 20:34:26 <sdake> any questions? 20:34:41 <sdake> if testing goes wrong, we may slip a day, so lets make sure testing goes right :) 20:34:55 <daneyon> sounds like a plan. 20:35:02 <sdake> ok folks thanks! 20:35:07 <jlabocki> thanks sdake 20:35:13 <rhallisey> thank you 20:35:14 <larsks> Thanks! 20:35:15 <rhallisey> :) 20:35:17 <sdake> thanks 20:35:18 <sdake> thanks :) 20:35:20 <sdake> #endmeeting