20:00:32 <sdake_> #startmeeting kolla
20:00:33 <openstack> Meeting started Mon Mar 30 20:00:32 2015 UTC and is due to finish in 60 minutes.  The chair is sdake_. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:34 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:36 <openstack> The meeting name has been set to 'kolla'
20:00:40 <sdake_> #topic rollcall
20:00:44 <daneyon_> hello
20:00:45 <jpeeler> hi
20:00:49 <sdake_> hey foolks sdake here ;)
20:01:20 <sdake_> samyaple around?
20:01:26 <sdake_> rhallisey slower?
20:02:04 <ccrouch> hiya
20:02:32 <rhallisey> hi
20:02:35 <sdake_> #topic ansible support in kolla
20:03:25 <sdake_> #link https://review.openstack.org/#/c/168637/
20:03:45 <sdake_> Sam has produced an initial patch which brings up mysql/mariadb
20:04:08 <sdake_> curious what folks opinions are into expanding our scope to handle deployment as well
20:04:12 <sdake_> eg, go a layer up
20:04:19 <sdake_> ie rather :)
20:05:11 <sdake_> don't all opine at once
20:05:13 <jpeeler> somebody had to do it
20:05:55 <jpeeler> might as well be us, especially with the way it's being done. it's not a requirement anyway, right?
20:06:19 <sdake_> which is not a requirement
20:06:22 <ccrouch> i don't think we should limit the number orchestration tools people want to layer around the basic containers
20:06:26 <sdake_> to use the ansible deployment models?
20:06:27 <jpeeler> the ansible usage
20:06:39 <sdake_> ya its optional
20:06:54 <sdake_> ccrouch you think introducing a optional deployment tool limits that number?
20:07:07 <jpeeler> right, i think great care should be taken to keep it optional
20:07:30 <ccrouch> sdake_ : no, i'm +1
20:07:44 <sdake_> my take is it will be optional if and until its the only way kolla is used ;)
20:08:09 <sdake_> daneyon rhallisey any thoughts?
20:08:19 <Slower_> do we want to more tightly define the kolla project to only be containers?
20:08:32 <Slower_> or do we want to have an optional install along with it..
20:08:35 <sdake_> slower_ that is the question on the table
20:08:38 <Slower_> I don't really care, just bringing it up that way
20:08:39 <ccrouch> sdake_ I'm saying: compose ~ ansible  ~ k8s
20:08:39 <ccrouch> the core team just has to support one model, to validate  the containers work
20:08:48 <Slower_> right
20:08:51 <Slower_> that makes sense
20:08:55 <Slower_> it would be helpful for dev
20:09:02 <ccrouch> and we shouldn't be building stuff into the containers that prevent other tools from being used
20:09:02 <sdake_> to really do ha, we need an orchestratoin tool (such as ansible)
20:09:32 <sdake_> ccrouch +1
20:09:43 <rhallisey> fine with me
20:10:02 <sdake_> ok folks well if you are into that review and your a core, just add a +2, and if there is a quorum of +2, I'll +a it
20:10:07 <sdake_> since its a scope expansion
20:10:14 <daneyon_> i think the heat template works for dev.
20:10:19 <ccrouch> sdake_: sure if you want extra QoS for your containers (i.e. HA) then you need a "better" tool than compose
20:10:36 <sdake_> ccrouch compose is good inbetweenansible and raw docker
20:10:40 <sdake_> but need something else
20:10:41 <ccrouch> +1
20:11:48 <sdake_> don't all cast your +2s at once :)
20:12:10 <Slower_> +1 :)
20:12:12 <sdake_> I guess we can come back to that - if your in favor of scope expansion please vote on that review
20:12:27 <sdake_> #topic k4 status
20:12:50 <sdake_> it looks from the launchpad not much work has been done
20:13:32 <sdake_> https://blueprints.launchpad.net/kolla/milestone-4
20:14:02 <sdake_> we have some outstanding blueprints
20:14:04 <sdake_> remember deadline is april 30th
20:14:24 <sdake_> anyone up for https://blueprints.launchpad.net/kolla/+spec/swift-container
20:14:46 <sdake_> i'd like storage to be a feature of 2015.1.0 ;)
20:15:20 <sdake_> slower you available to take that blueprint?
20:15:56 <sdake_> rhallisey https://blueprints.launchpad.net/kolla/+spec/cinder-container assigned to you in essentail state - not started
20:15:58 <Slower> I don't think I can commit to that right now, sorry
20:16:02 <sdake_> any progress on that?
20:16:04 <sdake_> thanks slower
20:16:27 <rhallisey> sdake_, haven't started yet
20:16:29 * Slower is slow
20:16:40 <rhallisey> I will get around to it this week
20:16:40 <sdake_> slower hence the name ;)
20:16:58 <sdake_> rhallisey if you could get cracking on that - that would rock - we need block storage for 2015.1.0
20:17:14 <daneyon_> sdake_updating my M4 bp. Should be accurate in LP in the next 10 min.
20:17:47 <sdake_> rhallisey one more essential assigned to you https://blueprints.launchpad.net/kolla/+spec/haproxy-container
20:18:09 <sdake_> marked not started atm
20:18:14 <rhallisey> same status.. Ill get around to cinder first
20:18:26 <sdake_> cool cinder is definately the priority to tackle first
20:18:27 <sdake_> thanks :)
20:18:32 <rhallisey> sure
20:18:46 <sdake_> somewhere in there we need a working ceilometer
20:18:50 <sdake_> there isn't a blueprint for that
20:18:58 <sdake_> ccrouch has it working from what I understand
20:19:22 <sdake_> ccrouch if you dont have time to git review the patches, please send me a diff and I'll submit them
20:19:33 <ccrouch> central agent, api and collector
20:19:45 <ccrouch> no alarming
20:20:02 <sdake_> well send what you have and ill make a blueprint for it
20:20:04 <sdake_> or did you want to tackle it?
20:21:02 <sdake_> just to keep in mind priorities for rest of kilo ha > storage > ansible
20:21:09 <sdake_> in that order :)
20:21:14 <sdake_> #topic continuous integration
20:21:27 <sdake_> so I had a longchat in #opnestack-infra with the infra team
20:22:02 <sdake_> that led to this review https://review.openstack.org/#/c/168598/
20:22:08 <sdake_> infra team wants us to add gates via tox
20:22:16 <sdake_> in summary
20:22:26 <sdake_> that review does that (incompelte atm, why its marked WIP)
20:22:41 <sdake_> but you will be able to do sudo tox -e functional
20:22:47 <sdake_> and bam, out comes the functional results
20:22:56 <sdake_> we should probably use nosetests to run the actual test cases
20:23:10 <sdake_> jpeeler you up to taking over that review work?
20:23:14 <sdake_> or want me to finish it off
20:23:22 <ccrouch> i thought openstack had dumped nose?
20:23:25 <jpeeler> i plan on finishing it
20:23:28 <ccrouch> for testr?
20:23:29 <jpeeler> and yeah, testr
20:23:33 <sdake_> sorry testr
20:23:37 <sdake_> same thing imo :)
20:23:44 <ccrouch> no comment :-)
20:23:51 <sdake_> jpeeler cool will I'd start with my review then
20:23:56 <sdake_> so your work is integrated in tox
20:24:09 <sdake_> rather then the pre-commit-hook
20:24:18 <sdake_> apparently pre-commit-hook is for git pre-commithook setups
20:24:22 <sdake_> not functional testing :)
20:24:32 <jpeeler> sdake_: i may have some more questions as i get to it
20:24:39 <sdake_> sounds good
20:24:55 <sdake_> #topic open discussion
20:25:03 <sdake_> there are a slew of outstanding patches
20:25:11 <sdake_> if folks coudl hit the review queue that would rock :)
20:25:28 <daneyon_> i'll def spend time later today on reviews
20:25:46 <rhallisey> same
20:26:07 <sdake_> how is he kolla on atomic effort going slower
20:26:38 <rhallisey> sdake_, we got it working
20:26:46 <rhallisey> like 20min ago
20:26:49 <sdake_> should submit a tools/atomic patch :)
20:26:55 <sdake_> or tools/start-atomic
20:26:58 <sdake_> or something like that
20:26:58 <rhallisey> ya will do
20:27:18 <sdake_> hopefully you used docker --env operation
20:27:23 <sdake_> rather then -e -e -e -e -e -e :)
20:27:39 <rhallisey> actually with atomic it looks like : atomic install glance
20:27:48 <rhallisey> and that's it
20:27:55 <sdake_> what about the metadata?
20:27:57 <rhallisey> we'll push those images soon so you can check em out
20:28:13 <jpeeler> rhallisey: how big are the images? (just curious)
20:28:17 <sdake_> where dot hthe config come from
20:28:29 <rhallisey> basically the images look like...
20:28:56 <rhallisey> http://paste.fedoraproject.org/204965/74732614/
20:29:21 <rhallisey> config come from one config file
20:29:41 <sdake_> how does /etc/openstack.env get into the image?
20:29:45 <rhallisey> which we setup using heat templates
20:30:01 <rhallisey> a heat template rather
20:30:43 <sdake_> you will need to make a build script I suspect to build those imgaes
20:30:54 <rhallisey> https://github.com/rthallisey/atomic-osp-installer/blob/master/heat/openstack_deploy.yaml
20:30:55 <sdake_> so people don't hae to manually docke rbuild
20:30:57 <rhallisey> ^ heat template
20:31:02 <sdake_> or better yet modify the exsting one
20:31:27 <rhallisey> right
20:31:56 <sdake_> interesting, not using the docker-compose integration point
20:32:05 <sdake_> but that should work for tripleo in the atomic-only cse
20:32:07 <rhallisey> we we're just testing with atomic
20:32:21 <rhallisey> there is also a docker-compose hook
20:32:26 <rhallisey> you can do either
20:32:44 <sdake_> interesting
20:33:04 <sdake_> well thats all the open discussion I wanted to have, any other folks have anything to chat about?
20:34:12 <sdake_> ok well then I'll end the meeting
20:34:23 <sdake_> for those lurking on these meetings later, join #kolla ;)
20:34:24 <jpeeler> sdake_: will the upcoming 1.6 release of docker contain all the fixes that kolla needs?
20:34:28 <sdake_> tia
20:34:34 <sdake_> jpeeler it is unclear
20:34:40 <sdake_> lk4d4 isn't sure if his stuff will be merged
20:34:46 <sdake_> which means docker locking is busted in 1.6
20:34:50 <jpeeler> wow, okay
20:34:51 <rhallisey> :(
20:34:56 <sdake_> I think the first version of docker that will work properly is 1.7
20:35:05 <sdake_> all prior versions will be busted for cmplex environments
20:35:37 <sdake_> in other words, starting and stopping more then one container makes docker bust :)
20:36:05 <sdake_> docker = rocket science
20:36:08 <sdake_> rockets blow up
20:36:12 <sdake_> #endmeeting