17:01:14 <pballand> #startmeeting CongressTeamMeeting
17:01:15 <openstack> Meeting started Tue Sep  2 17:01:14 2014 UTC and is due to finish in 60 minutes.  The chair is pballand. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:01:16 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:01:17 <thinrichs1> Hi all
17:01:18 <openstack> The meeting name has been set to 'congressteammeeting'
17:01:19 <pballand> hello all
17:01:22 <harrisonkelly> hey
17:02:21 <pballand> Lots of stuff happening last week
17:02:44 <pballand> we hit our alpha milestone
17:02:51 <pballand> thinrichs1: you mind giving an update of that?
17:02:58 <thinrichs1> pballand: sure.
17:03:15 <arosen> hi!
17:03:20 <thinrichs1> The alpha is ready for everyone to test.
17:03:34 <thinrichs1> The README has install instructions for both devstack and standalone.
17:03:45 <thinrichs1> You can write policy over Nova and Neutron.
17:03:50 <thinrichs1> And then monitor that policy.
17:04:01 <thinrichs1> The docs are pretty good.
17:04:08 <thinrichs1> There's a tutorial and a troubleshooting guide as well.
17:04:31 <thinrichs1> README contains info about how to view the docs (we're using RST, which you can compile to HTML).
17:04:51 <thinrichs1> File bugs as you find them, if you would.
17:05:13 <thinrichs1> We'll be sending an announcement to the ML later today.
17:05:42 <thinrichs1> I think the upshot of this is that everyone can now do an end-to-end trace through the (working) code.
17:05:52 <sarob> morning
17:06:08 <thinrichs1> I think that's about it for the alpha.
17:06:09 <pballand> thanks to all who helped on the alpha
17:06:21 <pballand> thanks thinrichs1
17:06:48 <pballand> to echo thinrichs1, if anyone has not yet tried the alpha, please do so
17:07:52 <pballand> harrisonkelly: any updates on what you’ve been working on?
17:08:15 <harrisonkelly> yes
17:08:30 <harrisonkelly> Derick suggested that I do the eventlet switch a different way, so I made minor changes
17:08:47 <harrisonkelly> but that introduced that weird error from run_tests.sh that I told you about before, the missing inbox
17:09:24 <harrisonkelly> the fix for that was to add a start() method to the deepsix class, but I cannot do that because of how Derick and me rewrote the deepsix class
17:09:51 <harrisonkelly> I’ve been off from work for a few days, so I haven’t had a lot of time to look into why thats happening yet
17:10:25 <pballand> would it make sense to push the change you have as an incremental change in the right diretion (away from threads)?
17:10:55 <harrisonkelly> the one that I have waiting for review now works, so I was waiting to see if I should push it or wait until it works because I changed a lot of files
17:11:35 <pballand> what patch are you referring to?
17:11:41 <harrisonkelly> one sec
17:11:50 <harrisonkelly> https://review.openstack.org/#/c/116308/
17:12:12 <harrisonkelly> that patch works (it ran run_tests.sh and tox -epy27 without errors)
17:12:12 <pballand> if you don’t want that pushed, you had better mark it as such
17:12:29 <pballand> (place a -1 on the workflow)
17:12:56 <pballand> that patch seems like a reasonable step in the right direction to me
17:13:04 <harrisonkelly> Should I do that? Or let it be merged (if it gets a check for the review) and then merge the new one in after so we can get rid of threading quicker
17:14:13 <pballand> it’s your code, what do you think?
17:15:04 <harrisonkelly> I think the patch I linked is step 1 to a two step process.  Step 1 removes threading and uses eventlets, step 2 adds more power to the eventlets and according to Derick, adds more power to the eventlets
17:15:40 <pballand> in that case, letting the patch go in makes sense to me
17:16:09 <pballand> (the faster you push incremental code, the lower the chance it will conflict with work others are doing)
17:16:28 <harrisonkelly> Yeah, I’m going to try to figure out whats causing it
17:16:43 <harrisonkelly> I think its because I’m automatically starting the eventlet when you create the object
17:17:13 <harrisonkelly> which is because you spawn the eventlet to create it, but if I had a start() method, it doesn’t start it until you explicitly say so
17:18:16 <pballand> harrisonkelly: ok - so the exiting code in review works, and you’ve got a couple kinks to work out in your next patch, right?
17:18:26 <harrisonkelly> correct
17:18:59 <pballand> harrisonkelly: sounds good
17:19:53 <pballand> anyone else have code updates to discuss before we move on?
17:20:59 <pballand> sarob: would you mind giving an update on the policy summit?
17:21:23 <arosen> I've been working on the congressclient hopefully I can get that out by the end of the week.
17:21:42 <arosen> That's all from me though. (I know i've been moving slow on the client) :/
17:22:24 <pballand> arosen: that will be great, thanks for the update
17:22:49 <arosen> also I have this patch here which changes the default port number: https://review.openstack.org/#/c/118207/ We might want to merge this before we announce the alpha
17:23:09 <arosen> but not super important either way probably.
17:24:14 <pballand> arosen: I like the port # change :)
17:24:39 <pballand> we already tagged the alpha, but it may be nice to get it in before most people pull (assuming they pull from master)
17:25:27 <pballand> hi rajdeep
17:25:58 <rajdeep> hi pballand
17:26:14 <pballand> rajdeep: we were just going through updates - do you have any updates to report?
17:27:14 <rajdeep> ok -- this week i worked on fixing the test cases
17:27:55 <rajdeep> for CollectionHandler webservice
17:29:49 <rajdeep> thats it from side
17:29:57 <pballand> arosen: pushed a workaround this morning for a bug that broke the gate
17:30:06 <pballand> https://bugs.launchpad.net/congress/+bug/1364515
17:30:26 <pballand> rajdeep: this looks like something you would be familiar with - would you be able to take a look?
17:31:26 <rajdeep> i remember that - perhaps replace the FakeClient with local Mocks
17:32:06 <thinrichs1> rajdeep: I have a few changes in review that do some cleanup of the Nova/Neutron clients.
17:32:16 <thinrichs1> It would be good if you could make your changes on top of those.
17:32:38 <pballand> rajdeep: replacing with local Mocks makes sense to me
17:32:46 <arosen> rajdeep:  yea I think local mocks is what you want to use. Or copy the fakes used in novaclient to congress.
17:32:47 <thinrichs1> Here's the last one:
17:32:48 <thinrichs1> https://review.openstack.org/#/c/118400/
17:36:03 <pballand> ok - I want to remind people that the policy summit is coming up
17:36:11 <pballand> #link https://etherpad.openstack.org/p/juno-midcycle-policy-summit
17:36:51 <pballand> we’ll be working on the schedule this week
17:37:48 <pballand> if you have a particular topic to discuss, please update the etherpad
17:37:58 <sarob> im back online when ready for me
17:38:15 <pballand> sarob: was hoping you could comment on the policy summit
17:38:20 <pballand> (or any other updates)
17:38:23 <sarob> yup
17:39:00 <sarob> so mikal will be asking a local to step in representing nova
17:39:31 <sarob> and writing up what nova policy looks like for him
17:40:02 <sarob> im going to get the gbp and heat guys to dump a bit into the etherpad as well this week
17:40:19 <sarob> we have about 60 people signed up right now
17:40:34 <sarob> almost 20 remotes
17:41:12 <pballand> should be good
17:41:19 <sarob> id ask anyone that has some interest in the use case discussion to add some thoughts to the etherpad as well
17:42:08 <sarob> thats it
17:42:46 <pballand> #topic Open Discussion
17:42:50 <pballand> thanks sarob
17:43:04 <pballand> that’s all I had for this week
17:43:12 <harrisonkelly> pballand: I’m not going to be able to make it to the meetings anymore because I have a class that falls right over this time block
17:43:46 <harrisonkelly> I can email you a status update every week, or let you know what I’ve been doing on the Congress channel if you’re on when I am
17:44:01 <pballand> harrisonkelly: that’s too bad - we hope you can still be involved
17:44:20 <pballand> harrisonkelly: you can absolutely keep in touch through the #congress channel and openstack-dev mailing lists
17:44:22 <harrisonkelly> I can still help, just not make the meetings
17:44:47 <pballand> harrisonkelly: good to hear - hope it’s a good class ;-)
17:45:01 <harrisonkelly> it’s GUI programming I, its a senior project class so I’m hoping its a good one
17:45:31 <pballand> harrisonkelly: sounds like fun :-)
17:45:53 <sarob> so about working on new data sources
17:46:56 <pballand> sarob: thanks for the reminder
17:47:31 <pballand> we are hoping to get people to start working on more data source plugins
17:47:51 <pballand> as part of that, thinrichs1 posted patches cleaning up the existing implementations
17:48:33 <pballand> we would like to recommend that everyone involved with the project writes a datasource driver
17:49:11 <pballand> doing so will help the project greatly, and also give each developer a good feel for how congress works
17:49:15 <arosen> I agree, i look forward to writing a datasource driver that integrates with something outside of openstack
17:49:43 * pballand needs to write a datasource driver as well :)
17:49:52 <rajdeep> perhaps we could list the drivers we would like to write
17:50:05 <rajdeep> then people could volunteer
17:50:41 <pballand> rajdeep: we could do that, but field is really wide open at this point
17:50:56 <pballand> any core OpenStack project is up for grabs
17:51:12 <rajdeep> ok i could write 1-2
17:51:32 <rajdeep> starting with keystone
17:51:33 <pballand> and there are numerous outside projects that would also make sense
17:51:41 <pballand> rajdeep: that would be great
17:51:58 <rajdeep> kudva wanted to write one for ceilometer
17:52:27 <pballand> if anyone is having trouble coming up with a component that needs a datasource driver, we can help them find something
17:52:52 <pballand> kudva and skn were both working on drivers, but they aren’t here this week
17:52:52 <rajdeep> there are other components we could write for - nsx, vsphere , open daylight outside openstack
17:53:28 <rajdeep> also cinder driver is up for grabs
17:53:46 <pballand> rajdeep: yes, the list goes on and on (Active Directory, antivirus, storage arrays, HVAC systems, ...)
17:54:02 <sarob> aws, gce, euca will be good ones once we knock out the major openstack projects
17:54:55 <rajdeep> i remember companies like tibco having a driver factory team at infosys - 100+ people
17:55:04 <rajdeep> with 200+ dirvers
17:55:31 <thinrichs1> The goal here is partly to build drivers but it's mainly to get people comfortable with Congress.
17:55:53 <thinrichs1> There may be a better way to integrate with many data sources.
17:56:08 <thinrichs1> For example, lots of the translation code is boilerplate
17:56:32 <thinrichs1> and if we had a description of the return value for an API call, we could in principle generate that code automatically.
17:56:38 <thinrichs1> Or have a generic translation algorithm.
17:56:54 <thinrichs1> But even if we had that, it's still valuable to roll one by hand.
17:57:04 <thinrichs1> From an educational perspective.
17:57:05 <rajdeep> ok, driver part is easier i would like to learn more about plexxi part
17:57:25 <sarob> so we want to have some mentoring around this
17:58:05 <sarob> teamwork on what good data source driver behavior looks like
17:58:38 <thinrichs1> sarob: I just made an editing pass through the drivers we have so others have good examples to work from.
17:58:42 <thinrichs1> Seems like a good first step.
17:58:43 <pballand> time check: one minute left
17:59:02 <sarob> thinrichs1: agreed
17:59:43 <pballand> that’s it for this week, thanks everyone for coming
17:59:53 <thinrichs1> Thanks all!
17:59:55 <pballand> #endmeeting