18:01:41 <pballand> #startmeeting CongressTeamMeeting
18:01:42 <openstack> Meeting started Tue Feb 11 18:01:41 2014 UTC and is due to finish in 60 minutes.  The chair is pballand. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:01:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:01:45 <openstack> The meeting name has been set to 'congressteammeeting'
18:02:18 <pballand> So I didn't update the agenda, but there are a couple things
18:02:26 <pballand> 1) change meeting time
18:02:32 <pballand> 2) discuss design doc
18:02:38 <pballand> 3) open discussion
18:02:47 <pballand> #topic meeting time
18:03:08 <pballand> we have heard from a few people that the meeting time doesn't work for them
18:03:35 <rajdeep> yeah too late for my time zone in india
18:03:48 <pballand> I think a PST time would be better for thinrichs and I
18:04:10 <pballand> would 1 or 2 hours earlier help rajdeep?
18:04:21 <rajdeep> 1 hour early would be better
18:04:42 <rajdeep> i attend solum meeting 2 hours earlier than this meeting
18:04:59 <pballand> if we can fine a room available, that would work
18:05:40 <pballand> there is #openstack-meeting-3 now
18:06:23 <pballand> so 1700 UTC on ever other Tues?
18:07:01 <pballand> anyone?
18:07:04 <rajdeep> yes that will be better
18:07:08 <thinrichs> Works for me
18:07:22 <pballand> #agreed meeting will be changed to #openstack-meeting-3 at 1700UTC
18:07:48 <pballand> I will make the changes on the wiki
18:07:58 <pballand> #topic Congress design doc
18:08:14 <pballand> thinrichs has been working on the design doc
18:08:31 <pballand> do you mind talking a bit about what you have and what should be done next?
18:08:51 <thinrichs> #link https://docs.google.com/a/vmware.com/document/d/1f2xokl9Tc47aV67KEua0PBRz4jsdSDLXth7dYe-jz6Q/edit#
18:09:25 <thinrichs> We've got the basic structure and there's plenty of text describing the basic design.
18:09:38 <thinrichs> There are a number of holes highlighted in yellow that need to be filled in.
18:10:26 <thinrichs> Some of the holes are conceptual details and some are API details.
18:11:08 <thinrichs> Probably the most useful thing at this point is for everyone to scan it and make high-level comments.
18:11:25 <thinrichs> Are we missing important design issues?  Do we agree with the things that are there?  Are there things we're missing?
18:11:41 <rajdeep> is the doc editable?
18:12:06 <pballand> Not publicly
18:12:11 <pballand> (as of now)
18:12:24 <pballand> I can add you (or anyone else who wants to contribute)
18:13:02 <rajdeep> sure, i was planning to start documenting existing implementation
18:13:23 <rajdeep> how are API server and the lexer parser linked in the current design
18:13:34 <pballand> they aren't :-)
18:13:47 <rajdeep> ok :)
18:14:03 <rajdeep> i was not able to figure out..i am assuming they have to be linked
18:14:04 <pballand> actually, there are 2 different entry points
18:14:32 <pballand> one just runs the parser, the other launches the API in an eventlet loop, with some policy engine integration
18:14:50 <pballand> but I have some uncommitted code on this that I have to clean up and get in
18:14:57 <rajdeep> ok..
18:15:17 <rajdeep> so people create their policies offline
18:15:23 <rajdeep> and hook into policy engine
18:15:38 <thinrichs> There are a bunch of things we're planning to do over the next month to get the prototype ready for others to use/understand.
18:15:47 <pballand> currently, yes, but the plan is that they can set/update policies via the API
18:15:53 <thinrichs> Here's the list.  Let's add to it/modify it.
18:15:56 <rajdeep> exactly
18:16:03 <thinrichs> - Hook up Congress to Nova/Neutron data sources (ideally in a way that makes it easy to hook up other data sources so that if customers have a use case with another data source then we/they can add it in short order)
18:16:05 <rajdeep> one entry point for everything
18:16:13 <thinrichs> - Hook up Congress to Nova/Neutron actions (so that we can actually correct violations after they occur)
18:16:20 <thinrichs> - Do some scaling tests/minor performance improvements.  Don't want to hand them something that only works with 10 hosts.
18:16:28 <thinrichs> - Add some error handling.
18:16:33 <thinrichs> - Some basic user docs
18:16:58 <rajdeep> are you planning to use nova / neutron python clients
18:17:02 <rajdeep> or REST directly
18:17:31 <pballand> I was thinking REST so they are more loosely coupled
18:17:48 <pballand> but performance could be an issue...
18:17:51 <rajdeep> ok...
18:18:08 <rajdeep> any plans to integrate congress into devstack
18:18:22 <rajdeep> ?
18:18:44 <pballand> that's probably the best way for integration testing
18:18:54 <pballand> but it isn't on my short-list
18:19:03 <pballand> it would be great if someone can tackle this
18:20:10 <pballand> #topic open discussion
18:21:06 <rajdeep_> let us get the nova/neutron integration first
18:21:17 <rajdeep_> devstack comes into picture after that
18:21:28 <pballand> that sounds reasonable to me
18:21:48 <pballand> I'll try to get some of my old commits cleaned up and pushed
18:21:51 <rajdeep_> where will congress pickup uses/ groups from
18:21:58 <pballand> and documenting a simple workflow with API is high priority
18:21:59 <rajdeep_> users*
18:22:39 <pballand> I did a simple hack for Active Directory
18:22:51 <rajdeep_> ok,
18:22:55 <pballand> but we should also support keystone
18:23:02 <rajdeep_> for devstack keystone would be default
18:23:07 <pballand> yup
18:23:23 <thinrichs> The goal, though, is for these to be examples of where users could be imported.
18:23:24 <rajdeep_> and why are we using ovs in the current implementation
18:23:28 <rajdeep_> saw some imports
18:23:55 <pballand> because that was an easy way for demonizing Python :)
18:24:19 <pballand> I should use oslo
18:24:28 <pballand> but was in a hurry and used what I knew
18:24:38 <rajdeep_> yeah..i got confused
18:24:43 <rajdeep_> looking at ovw
18:24:45 <rajdeep_> ovs
18:25:11 <pballand> yeah, it's just the ova-python utility routines
18:26:12 <rajdeep> ah ok..
18:26:18 <pballand> anything else to discuss today?
18:26:38 <thinrichs> How about next steps for each of us?
18:26:46 <rajdeep> i will go through the doc and ask questions if required
18:26:52 <thinrichs> From the list above, I was planning on ...
18:27:08 <rajdeep> what is the best way of communication : mailing list
18:27:34 <thinrichs> - Add some error handling and do overall cleanup of the policy engine
18:27:41 <pballand> yes, openstack-dev is probably the best place
18:27:54 <rajdeep> i will explore keystone integration
18:28:21 <pballand> ok, great
18:29:17 <pballand> so Feb 25, at 1700UTC is the next meeting
18:29:44 <pballand> until then, questions on the mailing list
18:29:55 <rajdeep> ok,
18:30:18 <pballand> thanks everyone
18:30:21 <pballand> #endmeeting