17:01:51 <thinrichs> #startmeeting CongressTeamMeeting
17:01:53 <openstack> Meeting started Tue Nov 25 17:01:51 2014 UTC and is due to finish in 60 minutes.  The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:01:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:01:57 <openstack> The meeting name has been set to 'congressteammeeting'
17:02:04 <banix> hi
17:03:00 <rajdeep> hi
17:03:07 <thinrichs> rajdeep: isn't it really late your time?  A heroic effort to attend!
17:03:19 <thinrichs> banix: hi!
17:03:20 <rajdeep> yeah ..it is
17:03:37 <rajdeep> somehow made it this time
17:03:59 <arosen1> Hi
17:04:10 <thinrichs> I just heard from someone else saying this meeting is at an odd time for them.
17:04:37 <thinrichs> Once we get people signed up for blueprints, we'll look into rescheduling to make sure everyone who has signed up can make the meetings.
17:04:47 <thinrichs> If that's possible.
17:05:01 <thinrichs> Let's start with status updates.
17:05:03 <thinrichs> #topic status
17:05:21 <thinrichs> sarob: want to start?
17:06:08 <sarob> Sure
17:06:46 <sarob> We have a bunch of people that have expressed
17:07:09 <sarob> The desire to commit code from bp at the summit
17:07:19 <sarob> Which is awesome
17:07:39 <sarob> I'm going to work on getting most of them
17:07:54 <sarob> Attending this meet weekly
17:08:20 <sarob> And pushing code and reviews weekly
17:08:47 <sarob> Thinrichs anything for me?
17:09:18 <thinrichs> Of the people who said they would push code for the next cycle, how many of them have put their name down on a blueprint?
17:10:21 <arosen1> me :)
17:10:40 <sarob> We have 9 people assigned
17:10:52 <sarob> I think it's the right time
17:11:03 <sarob> To start assigning bp
17:11:24 <sarob> To those people who have committed
17:11:37 <sarob> To pushing the code
17:11:48 <sarob> I'll follow up with them
17:12:15 <thinrichs> Do we have timeframes as well so we can assign blueprints to milestones?
17:12:19 <sarob> And assigning milestones would be helpful
17:12:51 <sarob> Huawei and Plexxi have
17:13:12 <thinrichs> The people who sign up for the blueprint are the ones who can tell us when they can get them done.
17:13:32 <sarob> Right
17:14:00 <sarob> We have that with Plexxi, Huawei, hp, Mirantis so far
17:14:25 <sarob> If possible
17:14:36 <thinrichs> Got it.  And you're going to follow up with the others.
17:14:38 <sarob> We should bring up
17:14:54 <sarob> Which bp need to be done sooner
17:14:57 <sarob> Than others
17:15:11 <sarob> Thinrichs right
17:15:38 <sarob> A preference for which bp get done in kilo
17:15:42 <sarob> And when
17:15:49 <sarob> Is good info
17:15:58 <thinrichs> If I remember right there aren't so many dependencies.
17:16:18 <thinrichs> Hopefully today I'll go through the blueprints, mark the dependencies, and assign priorities.
17:16:28 <sarob> So maybe just thinking about how many patches
17:16:35 <sarob> Can be reviewed
17:16:36 <thinrichs> (My bad that didn't get done before this meeting.)
17:16:55 <sarob> And how much testing mods
17:17:05 <sarob> To support new
17:17:18 <sarob> Code
17:18:27 <thinrichs> Okay, so it sounds like we have some action items for getting blueprints assigned to people and to milestones.
17:18:51 <thinrichs> #action thinrichs will add dependencies, priorities to blueprints
17:19:15 <sarob> I'd like to do a first pass today if you have time
17:19:28 <sarob> 18 dec is looming large
17:19:32 <thinrichs> #action sarob will follow up with people who volunteered for work to get estimates for time
17:19:49 <thinrichs> sarob: I already volunteered to make a pass today.
17:20:24 <sarob> Sorry, I'm volunteering to help
17:20:34 <sarob> Wasn't clear
17:20:47 <thinrichs> sarob: got it.
17:21:23 <thinrichs> I'll let you know when I make a pass, and then you can take a look.
17:21:30 <sarob> Cool
17:21:33 <sarob> I'm done
17:21:39 <thinrichs> Thanks!
17:21:59 <thinrichs> banix: want to go next?
17:22:18 <thinrichs> I know that last time you asked for priorities on blueprints.
17:22:32 <thinrichs> And I didn't get that done in time. :(
17:22:56 <banix> thinrichs: we are still trying to figure out our priorities with kudva and see how best we can contribute
17:23:17 <thinrichs> I had a chat with kudva and rayv last week.
17:23:44 <thinrichs> It sounds like rayv is interested in working on policies that define different notions of isolation (e.g. network isolation, compute isolation, etc.)
17:24:17 <thinrichs> So it sounds like they've got at least one concrete thing to work on.
17:24:25 <thinrichs> I also heard from kudva that he's thinking about delegation.
17:24:45 <banix> thinrichs: right, beyond that we have had more discussions to see if we can have a more focused effort
17:25:16 <thinrichs> banix: that would be great!
17:25:44 <thinrichs> I'd love to see a group of people take ownership of some feature-set or functionality.
17:26:00 <thinrichs> Let us know if you'd like to brainstorm.
17:26:09 <thinrichs> Or if there's anything else we can do to help.
17:26:18 <banix> thinrichs: sounds good. We are a  bit slow but we are getting there :)
17:26:24 <thinrichs> :)
17:26:59 <thinrichs> rajdeep: how are things going for you?
17:27:31 <rajdeep> good almost done with horizon integration and reviews
17:27:47 <rajdeep> waiting for final approval from janet
17:28:01 <thinrichs> Great!  Could you remind us what functionality you're adding to Horizon?
17:28:19 <jwy> rajdeep: i will review it again today
17:29:12 <rajdeep> yes integrating datasources details - list of data sources, list of tables in each data source and tuples supported in each table
17:29:40 <rajdeep> all this will should in horizon panels
17:29:48 <rajdeep> show*
17:30:24 <thinrichs> That sounds great!  Another thing that would be nice to show in Horizon is the status for every datasource.
17:30:28 <thinrichs> Obviously a separate change.
17:30:58 <thinrichs> But the status is really useful when you're debugging because it shows the error, if there was one.
17:31:41 <rajdeep> ok when you say status it means it is available or not
17:31:53 <thinrichs> And we also have the schema for each datasource: the names of the columns for each table.
17:32:23 <thinrichs> The status API call returns a dictionary.
17:32:45 <thinrichs> I think there are 2 key/value pairs right now: last_updated_time and error.
17:33:29 <thinrichs> But that status will be extended over time, so we'd probably just want to display the status as a dictionary (maybe a 2-column HTML table where the left column is the name of the key and the right column is the value).
17:33:41 <jwy> thinrichs: there's a bug open to split the data source rows into columns https://bugs.launchpad.net/congress/+bug/1379565
17:34:08 <jwy> rajdeep: do you want to add the status info in your patch, or we can do it separately?
17:34:14 <rajdeep> ok will add that after this CL completes it will be another panel / table
17:34:44 <rajdeep> jwy : would prefer as a separate patch :)
17:35:06 <jwy> rajdeep: sure
17:35:07 <thinrichs> A separate patch seems good to me too.
17:35:09 <rajdeep> i will fix that bug also
17:35:32 <jwy> rajdeep: i had already started on it
17:35:42 <rajdeep> ok..
17:35:43 <thinrichs> jwy: I *think* the rules are formatted when they get returned from the API.
17:36:19 <jwy> thinrichs: yeah i noticed recently that the rules include newlines when output
17:36:33 <jwy> need to convert those to breaks
17:36:54 <thinrichs> Not sure if that's the ideal way to do it, but it's really easy to do that kind of pretty-printing on the server-side since it already parsed the rules.
17:37:00 <arosen1> hrm i think our api should probably not be doing that on the server imo
17:37:19 <arosen1> i think we should parse the rules but probably not insert the \n's
17:37:26 <thinrichs> arosen: then we'll end up writing a parser in the CLI.
17:37:54 <arosen1> We're just returning json thou?
17:38:00 <thinrichs> It's much easier to strip the \n than it is to add them.
17:38:10 <thinrichs> Yes, we're returning JSON, but rules are strings.
17:38:29 <arosen1> ah okay i see. Yea that's probably fine then sorry.
17:39:03 <thinrichs> rajdeep: anything else?
17:39:24 <rajdeep> thats it from my side
17:40:14 <arosen1> I can go next.
17:40:19 <thinrichs> Great!  Horizon is really important and it will become moreso in the not too distant future.
17:40:34 <thinrichs> arosen1: okay
17:40:35 <arosen1> So last week we merged a few patches that re-factored the datasource driver base class. One in particular https://review.openstack.org/#/c/135099/ .
17:40:51 <arosen1> This now adds a small requirement on datasource drivers in which they can call self.register_translator(translator). This method ensures that the translator is valid when the datasource driver is loaded. This logic was being handled in convert_obj() before.
17:40:52 <arosen1> 
17:41:14 <arosen1> In addition, it also allows us to remove the method get_translators() from each datasource driver which was done here: https://review.openstack.org/#/c/135471/ and this now handled via the base class.
17:41:23 <thinrichs> I missed something there.
17:41:40 <thinrichs> The register_translator is something that you must call for anything to work?
17:41:53 <thinrichs> Or…what happens if you don't call register_translator?
17:41:54 <arosen1> thinrichs:  no this is called in the __init__ of a datasource driver
17:42:16 <arosen1> thinrichs:  for example, here's what had to change for the neutron driver: https://review.openstack.org/#/c/135099/7/congress/datasources/neutron_driver.py
17:42:49 <arosen1> thinrichs:  this fixes several things. 1) self.state is populated by register_translator()
17:43:17 <arosen1> also now the base class will know the full schema of a datasource driver when it loads.
17:43:29 <thinrichs> What happens if we don't call register_translator in init?
17:43:44 <thinrichs> Will it still work but there's no validation performed?  Or will it just not work?
17:44:18 <arosen1> thinrichs: one would need to do the validation themselves then.
17:44:52 <arosen1> Doing it this way allows us to have the base class to do all the validation.
17:45:20 <arosen1> and the validation of the translators should only need to be done once on load which is done now if one class register_translator()
17:46:09 <arosen1> thinrichs: alternatively one can call  validate_translator() in the base class here (https://review.openstack.org/#/c/135099/7/congress/datasources/datasource_driver.py) if they don't want to call register() though I don't really see any reason why one wouldn't want to use this method.
17:46:37 <thinrichs> Got it.  Is it also correct to say that if we see that something is missing from the datasource schema that we may have forgotten to call this method?
17:47:20 <arosen1> thinrichs:  totally :) that was what happened here: https://review.openstack.org/#/c/136133/ with the servers table that jwy reported
17:48:04 <arosen1> adding tempest tests though for get_schema() will ensure we catch these though in the future.
17:48:17 <thinrichs> Maybe we should add a note to our troubleshooting guide.
17:48:29 <arosen1> thinrichs:  sure, i can push a patch that does that not a bad idea.
17:48:41 <thinrichs> arosen1: anything else?
17:48:43 <arosen1> i'm also planning on rev'ing our writting a datasource driver docs to include this.
17:48:49 <arosen1> nope that's it. Thanks
17:49:05 <thinrichs> alexsyip: were you also planning to revise the datasource driver docs?
17:49:21 <alexsyip> I did a bit.
17:49:24 <alexsyip> It’s on gerrit.
17:49:29 <alexsyip> I think I added aaron as a reviewer
17:49:34 <arosen1> ah cool i'll check those out.
17:49:42 <alexsyip> thanks
17:49:53 <thinrichs> alexsyip: anything to report?
17:50:16 <alexsyip> I did a bunch of code reviews.
17:50:43 <alexsyip> and I’ve been working on a bug related to a demo.
17:51:02 <alexsyip> I ran into some behavior I didn’t understand, maybe aaron can help me with that.
17:51:25 <alexsyip> When I restart congress, there are some policy rules that congress loads from somewhere.
17:51:38 <alexsyip> But I don’t know where they’re coming from, and I cannot delete them using the CLI tool.
17:52:07 <alexsyip> arosen1: Where does the snapshot live?  Is it in /etc/congress/snapshot ?
17:52:35 <thinrichs> I think this may be a longer conversation than we have time for.
17:52:53 <arosen1> alexsyip:  i think that code is stuff that should be removed now that we have a database.
17:52:54 <alexsyip> ok, lets move it to the congress irc.
17:52:59 <arosen1> sounds goood to me.
17:53:06 <thinrichs> That'd be great!
17:53:17 <thinrichs> I want to be sure we get through statuses for everyone.
17:53:32 <thinrichs> jwy: how are things going?
17:53:45 <jwy> all right
17:54:16 <thinrichs> jwy: Anything to report?
17:54:29 <jwy> finished the spec for policy creation in horizon, fixed formatting in the tutorial, made some changes to the congress-specs testing (which needs review btw :))
17:54:33 <jwy> did a few reviews
17:55:20 <thinrichs> Great!
17:55:23 <jwy> https://review.openstack.org/#/c/136537/
17:55:40 <jwy> that's it
17:55:44 <thinrichs> I've been out of action for a while.  I'll try to make a round of reviews today/tomorrow.
17:56:00 <thinrichs> Did I miss anyone?
17:57:18 <thinrichs> Okay 3 minutes for open discussion.
17:57:21 <thinrichs> #topic open discussion
17:58:34 <thinrichs> Radu_: anything to report this week?
17:59:32 <thinrichs> Let's call that the end.
17:59:36 <thinrichs> Thanks everyone!
17:59:40 <thinrichs> #endmeeting