16:00:37 #startmeeting kolla 16:00:38 Meeting started Wed Sep 7 16:00:37 2016 UTC and is due to finish in 60 minutes. The chair is inc0. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:41 The meeting name has been set to 'kolla' 16:00:47 #topic rollcall 16:00:53 show of hands please 16:00:55 \0/ 16:00:56 o/ 16:00:57 hi o/ 16:00:59 o/ 16:01:08 o/ 16:01:12 \m/ 16:01:18 o/ 16:01:40 ola 16:01:46 #topic announcements 16:02:08 rc1 is coming live next week, after that only bugfixed happen 16:02:24 we deprecate debian and fedora 16:02:33 we will probably also work on kolla-ansible split, I'll start new vote on ML 16:02:53 to double confirm that core team is on board 16:03:03 hi :) 16:03:12 any community announcements? 16:03:15 o/ 16:03:46 #topic rc1 16:04:10 ok so rc1 means hard feature freeze for us, after that no blueprint code will be merged, we'll move new features to Ocata 16:04:31 o/ 16:04:40 and I would be extremaly glad if we focus on stability, testing and validation of stuff like upgrades 16:05:39 o/ 16:05:43 since other projects will have their rc's already, we can validate upgrades and make sure they're perfect by Ocata, or at worst Ocata-1 16:05:55 any comments/questions? 16:06:37 do we have a test plan? 16:07:12 if not we should create one to be sure that everything is working like expected, at least for the core components 16:07:15 well, nothing concrete, we can work a bit later in this meeting to brainstorm cases we want to have tested 16:07:45 good call berendt, I'll make sure to spend time on this later on 16:08:11 we have much scenario deploy now. maybe we should assign those to someone directly to make sure it works. 16:08:44 i agree Jeffrey 16:08:50 Jeffrey4l_, I'd rather ask for volunteers, but we need more tests, mostly around upgrades I'd say 16:09:14 we promised that kolla will upgrade seamlessly and we should make sure that's the case:) 16:09:16 like ironic? new added roles. 16:09:42 they all need be well tested, if we can get more volunteers. 16:10:05 agree 16:10:08 we need functional tests for all roles and upgrade tests for all roles that exists in newton 16:10:22 on a distributions that are not deprecated 16:10:25 yes. more gate is best. 16:10:51 agree, something like RDO do: https://www.rdoproject.org/testday/newton/testedsetups3/ 16:10:52 I don't think we can bet on gates now, this should be one of first things in ocata 16:11:16 good resource egonzalez90, thanks 16:11:34 inc0, yes. let's focus on the current work. 16:11:50 prioritiy change of https://blueprints.launchpad.net/kolla/+spec/multinode-gate to high? 16:11:59 and goal to o release? 16:12:00 I would start by validating that every new service is present in reno release notes 16:12:11 after that we can use this to prep proper test plan 16:12:30 berendt, yes 16:12:31 issue with multinode gate is that we're constrained by infra 16:12:51 done 16:13:09 what about a 3rd party? would this be possible? 16:13:15 but let's not talk about gates now, we'll have proper session in summit afair 16:13:15 multinode should be possible now ( only one controller ) 16:13:26 berendt, if some gracious company donates hardware 16:14:00 some bigger companies are working with kolla, we should ask them 16:14:16 yeah, let's make it a point in summit 16:14:22 let's move on about the validating 16:14:26 agree 16:14:48 #link https://etherpad.openstack.org/p/kolla-n-validation 16:15:03 timeslot: 10min - till :25 16:15:56 so 1. collect all the service need to verify 2. work out some scenario to test 3. need some volunteer to token the scenario and report the test result. 16:16:01 inc0 what means "compute kit"? 16:16:21 be 16:16:55 berendt: "compute kit" means testing the code on OpenStack that boots and nova VM... it mainly tests nova, neutron and glance.. 16:17:10 berendt, what you would call basic openstack deployment 16:18:33 inc0: in this case are we specificly refering to the default kolla install when we say "compute kit" 16:19:00 sean-k-mooney, I'd say default + cinder 16:19:17 however don't let it stop you from creating different test scenerios 16:19:37 we need to test as much as we can, but this *has to work 100 out of 100* 16:20:57 t-5min 16:23:44 what about smaller features like fernet tokens? 16:23:56 it's not small 16:24:02 and yes, we need this test too 16:24:07 yes, not really :) 16:24:51 ok, time's up, we'll get back to it next week 16:24:59 as still RC1 is being merged 16:25:18 #topic summit sessions 16:25:19 i will try to prepare a test matrix over the weekend 16:25:36 thank you berendt 16:25:44 #link https://etherpad.openstack.org/p/kolla-O-summit-planning 16:26:18 I'll need to double check with sdake what's the status of this 16:26:29 i thought we finished the session planning last meeting? 16:26:35 but while we need to provide space requirements, we can change topics till last minute 16:26:53 yes, I just want us to take another look if anything is missing or wrong 16:28:46 ok, I guess that's on me to double check with him 16:28:56 #topic kolla-kubernetes 16:29:25 anything you want to talk about kfox1111 and others? (seems we don't have much kolla k8s folks around) 16:30:03 guess not 16:30:09 #topic open discussion 16:30:27 soo, any pressing issues you guys have? 16:30:51 this meeting is quite fast, huh? 16:31:08 not from my site, we already discussed the deprecation stuff before the meeting 16:31:11 yeah, attendance wasn't great 16:31:20 - - 16:31:22 but that's the true before rc's 16:31:27 for the k8s: I think we do not have much work for newton? 16:31:30 we all know what we want to push forward 16:31:42 just begining phase? 16:31:51 duonghq, k8s doesn't *really* follow our rc policy yet 16:31:58 we don't want to freeze or whatever code there 16:32:09 I knew that, 16:32:12 that will start next release I guess 16:32:25 will there be a kolla-k8s release? 16:32:28 Yeah. I mean, we’re kind abound by the other stabalization concerns given we’re joined at the hip. 16:32:53 inc0: i guess it does not make sense to follow the release-with-milestones-trailing process until they have there 1.0 release 16:32:55 IMO, a kolla-k8s release is not a good idea atm 16:33:02 berendt, I don't think kolla-k8s is ready for 1.0 yet, am I correct wirehead_ kfox1111 ? 16:33:06 Naw. 16:33:21 than we do not have to take care of tests for k8s.. 16:33:28 oh snap completely missed the meeting 16:33:30 release 0.1 version ? 16:33:39 lol sdake 16:33:43 where was the notification ;) 16:33:52 sorry - i am of a day on my schedule 16:33:53 are there any news about the report from the OSCI tests? 16:33:54 because of labor day 16:33:57 Well, I’m not a real big fan of having twice-yearly releases. 16:34:09 not a problem sdake, we're almost done anyway 16:34:23 wirehead_ we can release k8s whenever we want 16:34:31 wirehead_, well, tough luck, you guys are OpenStack, that means we'll follow this cycle 16:34:36 wirehead_ but once we cut 1.0.0 its 6 month releasecycle 16:34:45 At this stage of the game, we could tag a ‘0.1’ that represents the current ‘hey, folks Nova works’ but otherwise, not too useful. 16:34:45 after we're done with first one 16:34:49 Yeah. 16:34:54 inc0 we are on a a release__independent cycle 16:34:56 with k8s 16:35:09 yeah, until 1.0 we can do whatever 16:35:12 wirehead_ thta has already beeen done - but not with the nova stuff 16:35:18 but after 1.0 we will follow normal cycle 16:35:37 trailing cycle that is 16:35:51 well, same thing as kolla, so yes 16:35:52 trailing 16:36:14 at the end I would be super happy if we synch release of all our projects 16:36:24 otherwise we'll be in dependency hell 16:36:50 agree with inc0 16:36:51 inc0 on ecomment from above 16:37:00 occata-1 is not a deadline for us for newton 16:37:15 we need to finish by the release_cycle trailing schedule 16:37:26 which is 2 weeks after the server projeects release their x release 16:37:43 yeah, I agree, but what I meant is people will normally start upgrading after o-1 16:37:55 so we need to make damn sure it's all good by then 16:38:01 right 16:38:18 I think it's about 1month after release itself 16:38:26 so we'll have stable releases of services by then 16:38:55 ok, any other topic? 16:39:43 guess not, thank you folks:) 16:39:51 thank you 16:39:51 let's get back to making release great 16:40:03 thanks guys. 16:40:06 oh one request from my side 16:40:33 https://review.openstack.org/#/c/366230/ testing this including upgrade is super important 16:40:56 so if someone have some time in their disposal, I would greatly appreciate help 16:40:59 release cycles are 45 days 16:41:11 although ocata is a shortened schedule 16:41:26 inc0 Jeffrey4l_ already said he would test upgrade 16:41:26 is it? 16:41:30 it is 16:41:32 yes ocata will be 16 weeks or 4 months instead of 6 mounts 16:41:40 sdake, yeah, but this needs as much testing as possible 16:41:43 I'll test it too 16:41:47 cool 16:41:52 but this is data consistency thing, critical 16:42:20 ya. more test is better. 16:42:41 especially that both me and Jeffrey generally run ubuntu 16:42:49 Jeffrey4l_ runs centos 16:42:58 ok, that's cool 16:43:13 I'm not super familiar with selinux stuff and this is one thing that I'd be caucious about 16:43:15 inc0, you will run ubuntu? 16:43:22 yeah I'll test xenial 16:43:26 cool 16:43:34 allright 16:43:36 thanks! 16:43:42 #endmeeting kolla