16:00:45 <sdake> #startmeeting containers
16:00:46 <openstack> Meeting started Tue Nov 25 16:00:45 2014 UTC and is due to finish in 60 minutes.  The chair is sdake. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:49 <sdake> #topic rollcall
16:00:50 <openstack> The meeting name has been set to 'containers'
16:00:51 <sdake> hey folks :)
16:00:52 <Slower> woo, good timing
16:00:59 <Slower> Ian Main here
16:01:01 <sdake> adrian is on pto this week and asked me to fill in as meeting chair this week
16:01:09 <daneyon> here
16:01:11 <yuanying-alt> o/
16:01:19 <sdake> most peopeol call me "steak" as in where is the beef
16:01:23 <diga> HiHi
16:01:42 <diga> Digambar Patil
16:01:56 <dguryanov> Hi!
16:02:00 <dguryanov> Dmitry Guryanov
16:02:05 <sdake> #link agenda -> https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2014-11-25_1600_UTC
16:02:18 <sdake> anyone have last minute items to add to agenda?
16:02:43 <sdake> #topic open discussion on initial architecture
16:03:06 <sdake> one approach to definining an initial architecture is to write it down - which has been done for magnum here:
16:03:37 <sdake> #link https://github.com/stackforge/magnum/blob/master/README.rst
16:03:53 <sdake> please take a few minutes to read through that, and then we can discuss if peopel like dislike etc
16:04:00 <sdake> and hone in on an architecture for our first milestone
16:04:54 <dguryanov> Are you planning to add networking other than docker's port mappings?
16:05:03 <sdake> ok, so there is my straw man, any questions?
16:05:14 <sdake> daneyon any thoughts on dguryanov's q?
16:05:30 <sdake> I think for our milestone 1, we aren't doing anything outside what is provided by base docker or kubernetes
16:05:33 <daneyon> sdake: I know we had some discussion last week around services. Is the plan to follow the existing k8s service model?
16:05:36 <sdake> the basic idea is to wrap docker & kubernetes
16:05:44 <sdake> with a rest API
16:05:59 <sdake> the wrapper would be generic enough that a "native" implementation could be written if one so desired
16:06:11 <sdake> but the basic idea is to reuse docker and k8s as much as possible
16:06:28 <sdake> this is just my proposal there may be others
16:06:34 <sdake> if there are others,  lets get em out on the table
16:06:58 <daneyon> there are several proposals in the docker community related to networking. 1 for adding to docker's native networking and others for creating a plugin model with network drivers.
16:06:59 <dims> o/
16:07:17 <dguryanov> So containers will use nova instance's IP. each pod will run in a separate net namespace, am I right?
16:07:31 <dguryanov> no, I'm wrong :)
16:07:36 <Slower> adrian otto has a different proposal
16:07:40 <Slower> I'm trying to find it
16:07:42 <dguryanov> All containers will run in the same net namespace
16:07:46 <daneyon> i think we need to keep things simple, using the existing k8s networking model until things flesh out (hopefully soon) in the docker community. thoughts?
16:08:17 <Slower> daneyon: I agree, from what I've seen of docker proposals they aren't getting a lot of traction
16:08:34 <Slower> seems like the core guys have ideas but I haven't seen them come to fruition yet
16:08:41 <Slower> I could be wrong but that's my impression
16:08:50 <sdake> docker and k8s are goin gto do what they are going to do
16:09:04 <sdake> nothing we do is really going  to be able to influence their direction at this point in time
16:09:13 * dims nods in agreement
16:09:16 <nitin_> sdake: what is k8s
16:09:23 <sdake> kubernetes
16:09:24 <dims> nitin_: kubernetes
16:09:26 <nitin_> Ok
16:09:28 <sdake> that is the magic that does pods/services
16:09:41 <nitin_> Have you guys started the containers meeting.
16:09:48 <sdake> yes we have nitin_
16:10:04 <nitin_> Ok
16:10:30 <nitin_> without roll call
16:10:36 <nitin_> seems to be
16:10:40 <sdake> nitin_ use the power of scrollback
16:10:44 <daneyon> Slower: I think traction will be made in the near future. Most of the proposals were created in the last few weeks.
16:11:24 <sdake> i prefer to focus on magnum proposals rather hten what docker and k8s are doing :)
16:11:45 <Slower> daneyon: ya agreed
16:11:47 <sdake> I'd like to flesh out a plan for what our dev will look like in the next couple weeks
16:12:06 <sdake> and set a milestone date
16:12:18 <Slower> nitin_: joined after rollcall
16:12:25 <sdake> anyone opposed to readme.rst serving as our intiial architecture?
16:12:28 <nitin_> yes
16:12:50 <dims> sdake: sounds good to me
16:12:51 <Slower> sdake: I like it
16:12:55 <diga> I am ok with this
16:13:08 <sdake> shall we have a vote then just to  make it official
16:13:10 <diga> sdake: +1
16:13:14 <sdake> +1 = like -1 don't like
16:13:18 <sdake> +1
16:13:20 <Slower> +1
16:13:20 <yuanying-alt> +1
16:13:25 <diga> +1
16:13:43 <sdake> #Link https://github.com/stackforge/magnum/blob/master/README.rst initial milestone #1 archite4cture
16:13:43 <daneyon> i think their is a strong possibility of this proposal being accepted for docker networking: # link https://github.com/docker/docker/issues/8997
16:13:54 <dims> +1
16:13:57 <nitin_> figuring out !
16:14:07 <sdake> ok sounds like we are set then with the initial milestone #1
16:14:08 <daneyon> The docker networking proposal uses this plugin model: #link https://github.com/docker/docker/pull/8968
16:14:17 <sdake> #topic adding of the blueprints
16:14:45 <daneyon> It may be a good idea for others to spend a few minutes to check out the proposals and chime in on the thread.
16:14:58 <sdake> #link https://blueprints.launchpad.net/magnum/milestone-1
16:15:56 <sdake> can we get a rollcall of blueprint ideas people are working on currently in irc?
16:16:28 <diga> yes
16:16:56 <sdake> just go ahead and type away - think of this as a brainstorm session :)
16:16:59 <dims> sdake: i was adding tests for the REST API
16:17:10 <diga> sdake: https://blueprints.launchpad.net/magnum/+spec/magnum-backend-docker
16:17:11 <dims> sdake: got 2 lined up
16:17:28 <diga> have you started this blueprint ?
16:17:29 <Slower> sdake: Are we going to run the magnum backend service in a container?
16:17:35 <Slower> eg, started by nova-docker
16:17:52 <sdake> slower I think that is too complex for milestone #1
16:17:59 <Slower> ok
16:18:07 <Slower> perfectly reasonable
16:18:18 <sdake> diga I have only implemented the binary
16:18:23 <diga> ok
16:18:26 <Slower> so for now just along side nova?
16:18:37 <sdake> there are two backends : https://blueprints.launchpad.net/magnum/+spec/magnum-backend-docker https://blueprints.launchpad.net/magnum/+spec/magnum-backend-kubernetes
16:18:42 <sdake> slower I think that is easiest
16:18:47 <Slower> agreed
16:18:51 <sdake> we can launch is something like heat-kubernetes heat template
16:18:58 <diga> Should I look at the remaining part if you allow me ?
16:19:00 <Slower> and from there it will be easy to explore vms/containers if we want
16:19:22 <sdake> any blueprint that doesn't have n assignee can have one - just click "pick me" diga
16:19:46 <sdake> slower can you file a blueprint tho, so that idea doesn't gert lost
16:20:00 <Slower> sdake: you bet
16:20:01 <sdake> dims can you file your 2 ideas lined up into launchpad and assign em to yourself
16:20:36 <sdake> we need someone to tackle https://blueprints.launchpad.net/magnum/+spec/db-object-definitions
16:20:48 <sdake> this connects into magnum-conductor
16:20:57 <sdake> anyone willing to take it on?
16:21:24 <yuanying-alt> o/
16:21:42 <sdake> yuangyin-alt is that a yes :)
16:21:51 <yuanying-alt> Yes yes
16:22:07 <sdake> https://blueprints.launchpad.net/magnum/+spec/db-object-definitions
16:22:14 <sdake> see if you can click "pick me" under assigned to
16:22:22 <sdake> rather "assignee"
16:22:36 <yuanying-alt> Ok
16:23:03 <sdake> diga were you planning tog take the docker backend or the kubernetes backend?
16:23:03 <dims> sdake: https://blueprints.launchpad.net/magnum/+spec/magnum-rest-api
16:24:12 <yuanying-alt> sdake: sorry I cant find "pick me".
16:24:30 <sdake> next to asignee isee the little circle?
16:24:46 <sdake> yellow circle with black pen in it i think :)
16:24:51 <sdake> try clicking that
16:25:07 <sdake> (sorry for not just setting I need  to see if foks can set their own blueprints)
16:25:23 <Slower> I'd be in for a docker or k8s backend
16:25:29 <Slower> or I can work with someone
16:25:37 <djsamp> I can help out with either of the two...
16:25:44 <dguryanov> I can also help with iot
16:25:54 <dguryanov> docker or kubernetes backeds
16:26:03 <Slower> sdake: what is the idea with doing both?  wouldn't it be easier to just pick one?
16:26:17 <sdake> slower docker backend g ives full access to container stuff
16:26:23 <sdake> k8s backend gives full access to k8s
16:26:27 <sdake> k8s hides alot of container tech
16:26:36 <diga> sorry, got disconnected
16:26:46 <sdake> diga you up for docker backend then?
16:26:54 <diga> yes
16:27:14 <diga> I am eager to do that
16:27:43 <nitin_> diga: I can help you with the docker backend.
16:27:46 <sdake> slower we can work on k8s together - i know alot of how it works
16:27:55 <Slower> sdake: ok sounds good
16:27:56 <sdake> nitin + diga = docker backend sounds good to me :)
16:28:17 <diga> +1
16:28:19 <diga> :)
16:28:22 <nitin_> +1
16:28:59 <sdake> yuanying-alt what is your alucnhpad id?
16:29:17 * dims liking all the forward progress :)
16:29:33 <yuanying-alt> sdake: launchpad id is "ootsuka"
16:29:54 <nitin_> diga: Can I also mention myself as assignee to the docker backend
16:30:07 <diga> Yes
16:30:07 <nitin_> my launchpad id is nitinagarwal3006
16:30:14 <diga> ok
16:30:19 <sdake> yuanying-alt do you mind taking all of the db stuff then?  Its probably all related
16:31:02 <yuanying-alt> OK.
16:31:09 <nitin_> diga: i can't see the pickme as assignee
16:32:10 <diga> nitin: i'll check it after the call
16:32:16 <dguryanov> I can help with some task this week
16:32:18 <nitin_> no problem.
16:32:37 <sdake> ok folks looks like all the blueprints that we are t racked are assigned
16:32:49 <sdake> are there other blueprints folks would like to see developed that are not tracked?
16:33:34 <Slower> looks good for m-1
16:33:37 <sdake> what about the client ?
16:33:48 <sdake> achandra is working on that, has anyone tried it lately?
16:33:51 <Slower> haha right
16:34:07 <sdake> I think we are blocked on getting a rest api call from the rest client to the backend to the conductor
16:34:11 <sdake> that is our next critical step :)
16:34:16 <Slower> yeah
16:34:21 <diga> yes
16:34:32 <sdake> dguryanov if you want o thelp on the client, that would rock
16:34:39 <sdake> since we are blocked on it atm
16:35:05 <sdake> https://github.com/stackforge/python-magnumclient
16:35:08 <dguryanov> OK, how can I contact achandra?
16:35:12 <sdake> ^^ for ot hers looking to unblock us
16:35:17 <sdake> try #openstack-containers on irc
16:35:30 <sdake> achanda is his nick sorry :)
16:35:45 <sdake> that project is in stackforge, so feel free to submit git reviews for it as well
16:35:52 <sdake> or contribute elsewhere as you see fit :)
16:36:06 <sdake> ok back on agenda :)
16:36:09 <dguryanov> OK, thanks!
16:36:11 <sdake> #topic initial mielstone deadline and objective
16:36:36 <sdake> we pretty much went through this already, but we have not set a deadline for when these blueprints would come together
16:36:50 <sdake> imo, these blueprints make up a fine milestone #1 (assuming they work) and the client works
16:37:09 <dims> before christmas? :)
16:37:21 <Slower> that's what I was thinking
16:37:26 <diga> yes
16:37:30 <sdake> yes
16:37:30 <Slower> try anyway :)
16:37:33 <sdake> 16th?
16:38:05 <sdake> that gives us 3 solid weeks of dev
16:38:20 <dims> +1
16:38:24 <Slower> or 19th..
16:38:27 <sdake> and 16-23 can be slip time if necessary
16:38:37 <dims> right
16:38:38 <Slower> yeah
16:38:44 <sdake> let sshoot for 16th and slip into 19th if we haver to
16:38:56 <sdake> ok, again lets vote on the deadline since we are held accountable ;)
16:38:59 * sdake flogs sdake
16:39:04 <dims> lol
16:39:18 <sdake> vote: deadline set for December 16th ,2014 for milestone #1
16:39:19 <dims> +1 for "shoot for 16th and slip into 19th"
16:39:24 <sdake> +1
16:39:27 <nitin_> +1
16:39:28 <yuanying-alt> +1
16:39:28 <diga> +1
16:39:29 <Slower> +1
16:39:33 <dguryanov> +1
16:39:51 <sdake> ok well that looks pretty much like a + 1 * 8 to me :)
16:39:57 <sdake> #topic open discussion
16:40:10 <sdake> cool well I'm really pumped we are making great progress
16:40:21 <sdake> we have 3 binaries written so far and now its just a  matter o ffilling them in
16:40:33 <sdake> (and the cleint of course)
16:40:38 <diga> yes
16:40:44 <sdake> On that note, any open discussion folks would like to have?
16:42:16 <sdake> ok well just a reminder we are pretty much blocked on client ->rest api comms atm -  I tried with curl and it wouldn't authenticate
16:42:25 <Slower> seems like there will be some API changing happening as we develop the backends especially
16:42:26 <sdake> so we need to sort that out - highest prtiority because i tblocks everyone's dev
16:42:41 <sdake> slower ack I agree the apis are just initial stabs
16:43:18 <sdake> ok well thanks for coming folks
16:43:21 <Slower> we'll have to work close to get that figured out.. especially trying to make it work for both k8s and docker
16:43:37 <sdake> next week we will also have a meeting - we can discuss christmsa downtime when adrian returns
16:43:45 <Slower> cool
16:43:47 <Slower> good meeting
16:43:53 <dims> thanks sdake!
16:43:56 <sdake> enjoy :)
16:43:58 <sdake> #endmeeting