17:03:34 <thinrichs> #startmeeting CongressTeamMeeting
17:03:35 <openstack> Meeting started Tue Oct 21 17:03:34 2014 UTC and is due to finish in 60 minutes.  The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:03:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:03:39 <openstack> The meeting name has been set to 'congressteammeeting'
17:03:53 <thinrichs> As usual, let's start with status updates.
17:03:59 <thinrichs> #topic statusUpdates
17:04:09 <thinrichs> sarob: want to start?
17:04:17 <sarob> sure
17:04:21 <rajdeep> hi
17:04:35 <thinrichs> rajdeep: hi
17:04:52 <sarob> our design session is scheduled for 03 nov monday 14:30-16:00
17:05:22 <sarob> i have asked if we can get it moved to tuesday to be co-located wih GBP
17:05:44 <sarob> does the team think this co-locating is a good idea or not?
17:06:14 <thinrichs> It'll work either way.
17:06:20 <sarob> okay
17:06:35 <sarob> i should know by mid week
17:06:51 <thinrichs> sarob: what should all of us do to prepare for the design session?
17:07:20 <sarob> there is an planning etherpad #link https://etherpad.openstack.org/p/par-kilo-congress-design-session
17:07:21 <s3wong> sarob: GBP is being place on Tuesday as part of "cross-project" slot, and given that Congress is on Monday, it is considered "Ops" session
17:07:45 <sarob> s3wong really?
17:08:30 <s3wong> sarob: I think that is the breakdown... at least according to ttx
17:08:58 <sarob> s3wong not sure that matters a bunch, thx for the info
17:09:46 <sarob> thinrichs i have listed out in the etherpad priorities, specs, bp, that have been on different discussions
17:10:03 <sarob> thinrichs we need to start paring it down
17:10:14 <sarob> thinrichs so we have a agenda
17:10:51 <sarob> thinrichs lining up bp to specs so they are one to one
17:11:11 <thinrichs> sarob: that looks good.  Maybe I'll take a pass over the etherpad and organize by topics
17:11:24 <thinrichs> e.g. UI, policy engine, data sources, etc.
17:11:37 <sarob> thinrichs the policy summit specs with at least pending patch link would be a good start
17:12:04 <sarob> thinrichs sure, i just did a big dump into the pad
17:12:23 <sarob> thinrichs im starting to clean it up some
17:12:40 <thinrichs> Then maybe we can organize the design session by prioritizing efforts within each group?  And then produce a total ordering over the most important features?
17:12:56 <sarob> thinrichs thats a good idea
17:13:34 <thinrichs> Before we all arrive at the summit, we should think through each of the groups of features to figure out what we each think is most important.
17:13:36 <sarob> i have a congress related summit sessions started at the bottom of the pad as well
17:13:55 <thinrichs> I'll send out an email reminder maybe a week before.
17:14:07 <sarob> thinrichs right on
17:14:39 <thinrichs> Any other prep we should be doing for the summit?
17:14:43 <sarob> i will list out the policy related sessions and talks that i have found in the etherpad bottom
17:14:54 <thinrichs> That'd be helpful!
17:15:02 <sarob> if anyone finds others pls add to the list
17:16:23 <sarob> i updated the agenda too
17:17:08 <sarob> not too crazy #link https://wiki.openstack.org/wiki/Meetings/Congress
17:17:53 <sarob> anything for me?
17:18:05 <thinrichs> Sounds good: do code reviews, look at outstanding bugs, review the etherpad for the summit.
17:18:37 <thinrichs> Anything else to report sarob?
17:18:59 <sarob> btw the congress session is under the ecosystem track
17:19:07 <sarob> thats all
17:19:21 <thinrichs> Thanks.
17:19:26 <thinrichs> alexsyip: want to go next?
17:19:39 <alexsyip> Yeah, the datasource driver code has gone through one review cycle with Tim.
17:19:57 <alexsyip> It’s all on gerrit now.
17:20:19 <thinrichs> It's good stuff.  I'd suggest you all take a look.
17:20:42 <alexsyip> The top level comment here https://review.openstack.org/#/c/125845/12/congress/datasources/datasource_driver.py
17:20:45 <thinrichs> Especially take a look at how the new neutron/nova drivers look.
17:21:07 <alexsyip> describes how someone would write a translator for a new data source driver.
17:21:32 <alexsyip> This is what an example driver would look like: https://review.openstack.org/#/c/127624/13/congress/datasources/neutron_driver.py
17:22:03 <alexsyip> This is the keystone driver, and it’s probably the simplest example: https://review.openstack.org/#/c/126731/13/congress/datasources/keystone_driver.py
17:22:38 <alexsyip> The neutron one is a little more involved because the neutron driver provides more data, in a more complex organization with several nested lists and dicts.
17:23:37 <thinrichs> So instead of writing Python code that translates the output of an API call to tables, you just provide a declarative description of the output of the API call.
17:23:44 <thinrichs> And the translation to tables is handled for you.
17:23:49 <rajdeep> alexsyip looks very simple
17:24:17 <LouisF> alexsyip: can these be used as a model for a gbp translator?
17:24:21 <thinrichs> alexsyip: do you expect this to work for every datasource driver?
17:24:33 <alexsyip> Thanks!  The example drivers show the declarative description, and a little bit more plumbing to connect that to the API client code.
17:24:55 <alexsyip> LouisF: do you mean to read data from GBP or to write data to GBP ?
17:25:12 <LouisF> alexsyip: either
17:25:32 <LouisF> alexsyip: or is it just read?
17:26:01 <alexsyip> thinrichs: I think this code will cover most data source drivers, but there could be some that don’t provide data in the same way (with lists and dicts).  For that, we may have to hand code those drivers, or convert the data into lists and dicts.
17:26:12 <alexsyip> LouisF: We should be able to use this when reading from GBP.
17:26:30 <alexsyip> This does not have any code to assist with writing to GBP right now.
17:26:49 <LouisF> alexsyip: great - what is needed to writew to gbp?
17:29:16 <alexsyip> LouisF: So when we read GBP data into congress, we’re just pulling in the GBP data as raw input.  It’s likely that the congress policies will need to do some further interpretation of the raw data.
17:29:38 <alexsyip> LouisF: for the purposes of writing policies in Congress.
17:30:29 <alexsyip> LouisF: For writing to GBP, our original plan was to have congress tables that trigger on row insert and delete.
17:30:42 <alexsyip> We don’t have that framework yet, nor any drivers of that type.
17:30:49 <LouisF> alexsyip: ok thx
17:31:11 <alexsyip> Ok, that’s all I have irght now.
17:31:19 <thinrichs> thanks alexsyip: good stuff!
17:31:38 <thinrichs> rajdeep: want to go next?
17:32:42 <rajdeep> sure
17:33:02 <rajdeep> i have been working on datasource integration with horizon
17:33:10 <rajdeep> panels and tabs are ready
17:33:38 <rajdeep> working on last leg to pass multiple values
17:34:08 <rajdeep> datasource and tables are in the first table
17:34:40 <rajdeep> for each row -- there is a child view with actual values
17:35:12 <thinrichs> UI is super-important.  I'm looking forward to seeing it in action!
17:35:23 <rajdeep> hopefully by next week i will have code completed
17:36:37 <rajdeep> that completes my status
17:36:42 <thinrichs> Sounds good.  Let us know if you need anything from us.
17:37:44 <thinrichs> LouisF: anything you'd like to report on?
17:38:13 <rajdeep> sure
17:39:08 <thinrichs> Maybe we'll come back to LouisF
17:39:19 <LouisF> thinrichs: no
17:39:42 <thinrichs> Sorry for jumping the gun there, LouisF.
17:40:12 <thinrichs> jasonsb, glebo, and I continued our discussion from last week in #congress after the meeting ended.
17:41:04 <thinrichs> glebo volunteered to summarize what we had discussed.
17:41:11 <thinrichs> jasonsb: have you heard from glebo?
17:41:13 <LouisF> nothing more than https://review.openstack.org/#/c/129350/
17:41:37 <jasonsb> thinrichs: no sir.   i don't have any contact info so i haven't reached out
17:42:14 <thinrichs> jasonsb: Okay.  I'll try to follow up with him today.
17:42:33 <jasonsb> sounds good
17:42:33 <thinrichs> LouisF: I'll look into that review today.
17:42:55 <thinrichs> Last I saw it looked to be in good shape.
17:43:26 <jasonsb> i need to look at rajdeep's work before i have anything coherent to add anyway
17:43:40 <thinrichs> jasonsb: sounds good.
17:43:51 <thinrichs> rajdeep: have you pushed anything to review that jasonsb could look at?
17:44:10 <thinrichs> rajdeep: or could you provide him with pointers to the existing UI work?
17:45:08 <rajdeep> thinrichs not yet
17:45:35 <rajdeep> will do so when i am ready with the first cut can send some screenshots for now
17:46:00 <jasonsb> that sounds godo
17:46:04 <rajdeep> jasonsb : please share your email details
17:46:12 <jwy> the existing ui code is under congress/contrib/horizon
17:46:49 <thinrichs> jwy: didn't recognize your IRC handle.
17:47:05 <thinrichs> jwy: welcome!
17:47:21 <jwy> the congress devstack script has been modified to set up horizon for congress
17:47:24 <rajdeep> jwy has been of great help on ui side
17:47:48 <jwy> thinrichs: thanks! i realize my handle here is different than my username at work :)
17:47:52 <thinrichs> rajdeep, jwy, jasonsb: Maybe the 3 of you can start up an email thread.
17:48:46 <thinrichs> jasonsb: I think we'd just need your email address (and if you'd rather not put it here in IRC, you can email me and I'll forward it.)
17:48:47 <jasonsb> sounds good
17:49:15 <jasonsb> jason.bishop@gmail.com
17:49:27 <thinrichs> jwy: anything else you'd like to report on?
17:49:29 <jasonsb> (so much easier)
17:50:22 <jwy> mainly been helping rajdeep, haven't gotten back to writing any new code myself since the last push
17:50:27 <jwy> planning on doing more this week
17:51:04 <thinrichs> Sounds good.  Let us know if you need anything from us.
17:51:33 <thinrichs> The new thing I started experimenting with this week is based on a request from the ML.
17:51:55 <thinrichs> Someone wanted to be able to analyze policy.json files to understand what roles were required to execute a given command.
17:52:16 <thinrichs> So I started writing the code to pull policy.json files into Datalog and do the analysis with Congress.
17:52:37 <thinrichs> I'm about done pulling the policy.json files in.
17:52:49 <thinrichs> My initial experiments with doing the analysis seemed promising.
17:53:24 <thinrichs> Not only is this a potentially useful bit of functionality for OpenStack, but it also gives us a concrete example of how we might start thinking about pulling in policy from other policy engines.
17:53:44 <thinrichs> Anyway, I've spent half a day on it so far, so it's too early to have much to report.
17:54:19 <kudva> I might've missed this discussion before: 1. Is there any update on Ceilometer integration (need/interest/plan/progress etc).
17:54:20 <thinrichs> Okay.  5 minutes left.  Time for open discussion.  (Or for someone I've missed to give a status update.)
17:54:48 <thinrichs> kudva: there's a ceilometer driver commited.
17:55:01 <kudva> thinrichs: who is working on it? I want to help out there
17:55:04 <thinrichs> From what I understand, it doesn't pull statistics out.
17:55:43 <thinrichs> pettori was working on it, but I see he's not on IRC right now.
17:56:00 <thinrichs> alexsyip was going to look into using ceilometer for a use case.
17:56:17 <kudva> thinrichs: which particular use case?
17:56:17 <thinrichs> Maybe the two of you can connect to figure out what needs to be done.
17:56:33 <kudva> thinrichs: yes, is alexsyip online?
17:56:56 <alexsyip> Yes
17:57:15 <kudva> alexsyip: let's connect offline, would be interested in hearing about your use cse
17:57:23 <alexsyip> ok
17:57:31 <Radu_> I would like to be able to pull data from the DSE instead of using API calls from a driver, is that feasible right now?
17:58:15 <thinrichs> You mean you want to pull data from a deepsix instance?
17:58:36 <alexsyip> kudva: You can email me at ayip@vmware.com
17:58:37 <thinrichs> Would that be an instance of DatasourceDriver?  Or the policy engine?
17:59:01 <Radu_> Policy
17:59:07 <thinrichs> Datasource drivers properly implement publish/subscribe.
17:59:17 <Radu_> ok
17:59:19 <thinrichs> The policy engine doesn't publish anything today.
17:59:27 <Radu_> Ah.
17:59:50 <Radu_> Ok thanks, I was looking around and saw some #todo's but hadnt dug into it too much yet
18:00:03 <sarob> remember everyone to review and update #link https://etherpad.openstack.org/p/par-kilo-congress-design-session with your kilo design session topics and use cases
18:00:20 <thinrichs> Radu_: let's continue on #congress.  I'd like to hear more about what you want/need.
18:00:25 <thinrichs> We're out of time.
18:00:29 <thinrichs> Thanks all!
18:00:30 <sarob> cheers
18:00:36 <thinrichs> #endmeeting