20:02:04 <sdake_> #startmeeting kolla
20:02:05 <openstack> Meeting started Mon Apr  6 20:02:04 2015 UTC and is due to finish in 60 minutes.  The chair is sdake_. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:02:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:02:09 <sdake_> #topic rollcall
20:02:10 <openstack> The meeting name has been set to 'kolla'
20:02:22 <sdake_> hey folks :)
20:02:28 <daneyon> hey
20:02:35 <rhallisey> hi
20:02:56 <sdake_> jpeeler making it today?
20:02:57 <jpeeler> hey
20:03:00 <sdake_> he ythere :)
20:03:10 <sdake_> slower alive?
20:03:39 <sdake_> #topic announcements
20:03:47 <sdake_> quick couple of announcements to make
20:04:16 <sdake_> first off, last I checked, the kolla talk at summit was #5 most popular talk on sched.org
20:04:31 <sdake_> so if your a core and your going, try to get there early - I'd like all the cores in the room if possible
20:04:54 <rhallisey> sure thing
20:05:04 <sdake_> i dont know what kind of room capacity they will give us
20:05:07 <sdake_> the room is already assigned
20:05:19 <sdake_> but there is about 65 people registered atm
20:05:45 <sdake_> second thing, upstream docker is at rc4, and I get the impression docker 1.6.0 is coming RSN
20:05:46 <sdake_> like this week
20:06:19 <sdake_> once docker 1.6.0 is out we can stop carrying our own binaries and whatnot
20:06:21 <sdake_> yay :)
20:06:31 <daneyon> woohoo
20:06:50 <sdake_> third thing, docker-compose PR for pid support has support from two reviewers and was awaiting integration test additions - which I just finished prior to this meeting
20:07:06 <sdake_> so prior to summit, Id expect we can run everything from master or released packages without any magic :)
20:07:07 <sdake_> yay
20:07:13 <sdake_> anyone else have any announcements?
20:07:15 <daneyon> nice
20:07:54 <sdake_> #topic k4 planning
20:08:06 <sdake_> I just noticed rhallisey cranked out some cinder containers
20:08:15 <sdake_> but haven't reviewed yet
20:08:17 <sdake_> yay cinder :)
20:08:23 <rhallisey> uh
20:08:25 <sdake_> rhallisey do they work ? :)
20:08:26 <rhallisey> its still iwp
20:08:28 <rhallisey> no
20:08:31 <sdake_> ok
20:08:38 <sdake_> well step in the right direction atleast :)
20:08:40 <rhallisey> i trieda na all inone
20:08:59 <rhallisey> tried to put them all in one
20:09:06 <rhallisey> to see if i could get something going
20:09:09 <rhallisey> didn't quite work
20:09:25 <daneyon> rhallisey i'm in at the same place with neutron agents
20:09:27 <sdake_> recommend reading the ops guide
20:09:31 <rhallisey> so I pulled them apart by service
20:09:46 <sdake_> it says step by step how to configure the services
20:09:51 <rhallisey> ok
20:10:02 <rhallisey> ya I have a config file that looks good
20:10:07 <rhallisey> but ill check it out
20:10:09 <daneyon> last night i started testing running multiple processes in a container using supervisor
20:10:10 <sdake_> danyeon you want to give us an update on neutron?
20:10:46 <sdake_> bdestur said l3 and dhcp need to share some part of the filesystem via a volume mount
20:11:05 <sdake_> are you having trouble there or elsewhere?
20:11:09 <daneyon> struggling with the agents containers. i pulled in a neutron dev, but still trying to figure things out. Last night, i started to consolidate the agents in a single container and use supervisor.
20:11:20 <daneyon> this is like the OSDA model
20:11:43 <daneyon> If that works, are others cool with this model at least for our first iteration of Neutron?
20:11:47 <sdake_> samyaple said he had a problem with netns not being visiable outside the container for the agents to manage
20:12:04 <sdake_> first iteration is good, but I'd like to not use supervisord if at all possible
20:12:15 <sdake_> but if it works together it will work separately
20:12:22 <sdake_> just a matter of finding out where the state is shared
20:12:28 <daneyon> ya, containers spawning net namespaces is far outside the typical containerized app behavior.
20:12:46 <sdake_> with --net=host it should just work (TM)
20:13:07 <daneyon> sdake_ Agreed,I want to get them to work together then look at fine tuning.
20:13:09 <sdake_> daneyon was neutron not working prior?
20:13:46 <daneyon> I would prefer if the combined agent container s is accepted for our initial Neutron release, so I can make progress with HA. Then come back to improving Neutron.
20:14:05 <daneyon> sdake_ I thought it was
20:14:07 <sdake_> that wfm, neutron is definately #1 priority atm, followed by ha
20:14:28 <daneyon> adake_ agreed. I want to get Neutron out the gate and then go to HA
20:14:32 <sdake_> if I demo nova networking at summit i'm going to get kicked out of the room :)
20:14:52 <daneyon> I think it comes down to what will we accept for our 1st iteration of Neutron.
20:15:07 <sdake_> fat container, with idea of making it skinny :)
20:15:17 <daneyon> combined container or does the 1s iteration need to have each agent running in seperate containers?
20:15:32 <sdake_> danyeon get whatevery oucan to work to work
20:15:33 <daneyon> i'm a fat container
20:15:39 <sdake_> and we can revisit in liberty
20:16:06 <daneyon> cool, that's what I thought and why i started down the supervisor road yesterday.
20:16:13 <daneyon> sweet
20:16:30 <sdake_> once you finish up, and get things working reliably, I might take a stab at making it skinny
20:16:32 <sdake_> time will tell :)
20:16:58 <sdake_> jpeeler any action on functional tests?
20:17:13 <daneyon> i may do the same, I would just really like to have multi-node HA for Vancouver. I know their is a lot of work ahead of us to make that happen.
20:17:13 <jpeeler> negative, but i haven't forgotten about them
20:17:44 <sdake_> daneyon ya I'm holding out hope we can get 3 node ha controller working, but if it isn't, thats ok with me
20:17:48 <sdake_> i'd rather we have the main services working
20:17:58 <sdake_> neutron being prime one :)
20:18:29 <daneyon> agreed. I'll definitely get Neutron working, it just may not be optimal for the time being.
20:18:33 <sdake_> jpeeler I am going to be submitting a review shorty that adds tox -e functional testing
20:18:39 <sdake_> but it doesn't actually run any test cases
20:18:55 <sdake_> there will be a tests/setup.py
20:18:58 <sdake_> a person would run that
20:19:02 <sdake_> then run tox -e functional
20:19:09 <sdake_> and the functional test cases would be run via testr
20:19:19 <sdake_> jpeeler sound good?  then you can integrate in that directory
20:19:30 <jpeeler> sdake_: sounds great, you've done a lot of the work
20:19:47 <sdake_> well I dont know how to write the test cases :)
20:19:51 <sdake_> so that will be a challenge
20:20:04 <sdake_> i think our first good test case should be start followed by doing some basic commands
20:20:18 <sdake_> i'm not quite sure how to do that from python
20:20:47 <jpeeler> i have an idea, so we'll see
20:20:51 <sdake_> nice
20:20:59 <sdake_> ok, well lets keep cracking at it
20:21:07 <sdake_> lots more peope are interested in kolla then 3 mo ago
20:21:16 <sdake_> so I expect we will have our hands full in iberty :)
20:21:24 <sdake_> #topic open discussion
20:21:31 <sdake_> nice short meeting - 21 minutes :)
20:21:36 <sdake_> anyone have any open topics to discuss
20:22:50 <sdake_> ok then
20:22:57 <sdake_> thanks for coming :)
20:23:00 <sdake_> #endmeeting