00:01:11 <thinrichs> #startmeeting CongressTeamMeeting
00:01:11 <openstack> Meeting started Thu Dec 24 00:01:11 2015 UTC and is due to finish in 60 minutes.  The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:01:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:01:15 <openstack> The meeting name has been set to 'congressteammeeting'
00:01:30 <thinrichs> Since it's so close to the holidays, who is here?
00:01:40 <ramineni> hi
00:01:44 <pballand> hi
00:01:44 <ekcs> hi
00:01:50 <thinrichs> ramineni, pballand, ekcs: hi
00:02:08 <masahito> hi
00:02:12 <thinrichs> masahito: hi
00:02:17 <thinrichs> I have a brief agenda today...
00:02:20 <thinrichs> 1. distributed arch
00:02:22 <thinrichs> 2. python3
00:02:27 <thinrichs> 3. Austin
00:02:37 <thinrichs> Anything else?
00:03:05 <thinrichs> So let's get to it.
00:03:10 <thinrichs> #topic distributed architecture
00:03:16 <thinrichs> pballand: how's dsenode progressing?
00:03:53 <pballand> lots of progress
00:04:03 <pballand> (typing)
00:05:06 <pballand> In addition to the skeleton already pushed to review, I’ve added the control bus back in - currently debugging an issue with control bus message deliver, but I’m pretty sure I have a handle on it
00:05:41 <pballand> I expect to push another revision tomorrow before taking some time off for the holiday
00:06:25 <pballand> featureset to include node that can run multiple services, and discover other nodes and their services; making rpc calls between nodes and services
00:06:42 <thinrichs> When do you expect to have something we can use to start hooking up the datasources/api/policy?
00:07:06 <pballand> next step is to tie in table pub/sub
00:07:22 <pballand> at which point we can hook up existing datasources
00:08:10 <pballand> not sure on timing as I’ve got limited time to work on this for the next 1.5 weeks
00:09:18 <thinrichs> Definitely enjoy your holiday!  But assuming we start the clock from Monday after New Years, how long?
00:09:27 <thinrichs> 1 week, 2 weeks, … ?
00:10:04 <pballand> 1 week to review, unknown how long reviews will take
00:11:31 <thinrichs> 1 week sounds great!  So Jan 13 we should all expect to start reviewing.
00:11:59 <thinrichs> (That's the next IRC 1 week after the new year.)
00:12:28 <thinrichs> That should give us just enough time to make a first attempt at hooking up some real datasources.
00:12:36 <thinrichs> Before the mid-cycle sprint.
00:12:47 <thinrichs> And then at the sprint, we can take a day to try to get everything working.
00:13:03 <thinrichs> Any other questions or comments?
00:13:40 <thinrichs> #topic python3
00:13:54 <thinrichs> We are now officially gating on python3!
00:14:05 <thinrichs> ekcs (and others): thanks for getting this done!
00:14:28 <ekcs> nice
00:14:33 <thinrichs> That's all on this from my end.
00:15:26 <ekcs> I plan to go through and add several import __future__
00:15:40 <ekcs> to make sure the behavior is consistent between 2 and 3 for future code added.
00:15:56 <ekcs> things like real_div and print_function
00:16:13 <thinrichs> ekcs: you know if that makes sense better than me.  Anyone else have comments?
00:17:22 <thinrichs> Moving on then...
00:17:24 <thinrichs> #topic Austin
00:17:25 <pballand> question
00:17:30 <thinrichs> shoot
00:17:51 <pballand> do we still support 2.x?  What’s the recommended dev environment (2.x, 3.x, both)?
00:19:16 <ekcs> not sure what the latest official position is. but looking at the specs and reviews it seems like the goal is to support both 3.4 and 2.7 in the mean time.
00:19:43 <thinrichs> Here's the relevant page..
00:19:45 <thinrichs> #link https://wiki.openstack.org/wiki/Python3
00:20:02 <thinrichs> ekcs: do you know if our python-client is python3 compatible?
00:20:20 <ramineni> thinrichs: it is
00:20:27 <thinrichs> ekcs: I see on that page python3 status for some of the openstack projects.  It'd be good if we could include congress on it.
00:20:28 <pballand> thanks for the link - is there a TL;DR section?
00:20:38 <ekcs> thinrichs: okay.
00:20:39 <ramineni> thinrichs: we gate python3 job there
00:21:21 <pballand> I guess it’s “make sure everything works in 2.7 and 3.4"
00:21:23 <pballand> correct?
00:21:41 <thinrichs> I see Congress listed.
00:21:50 <thinrichs> pballand: I think that's right.
00:22:10 <thinrichs> in Liberty they moved to supporting only 3.4 and 2.7, which to me means that's a safe bet for mitaka too.
00:22:50 <thinrichs> Does anyone know what the Python3 classifiers means?
00:22:54 <pballand> gotcha, thanks
00:24:43 <thinrichs> ekcs: I suspect some of the data from that page is generated from projects.yaml.  So before manually updating the Congress stuff, it's probably worth figuring out if they regenerate it ever.
00:25:08 <ekcs> thinrichs: Okay. In terms of status, what would you say our status is? python3 is there, but not really tested that much.
00:25:14 <thinrichs> ramineni: BTW thanks for confirming py3 on the client.
00:25:32 <ekcs> thinrichs: python3 compat: in progress; py3 voting: yes?
00:26:13 <thinrichs> This may be a poor approximation, but all the tests pass and we're voting.  I think that means we're python3 compatible.
00:26:38 <thinrichs> There may be bugs, but there can always be bugs.
00:27:16 <ekcs> thinrichs: okay.
00:27:33 <thinrichs> Moving on...
00:27:36 <thinrichs> #topic Austin
00:28:00 <thinrichs> The next summit is in Austin, and they've opened up the call-for-speakers.
00:28:01 <thinrichs> #link 	1.	https://www.openstack.org/summit-login/login?BackURL=/summit/austin-2016/call-for-speakers/
00:28:21 <thinrichs> One new thing this year is that they limit speakers to 3 submissions.
00:28:52 <thinrichs> I don't think any of us ever exceeded 3, but it's a sign that they're getting too many submissions.
00:29:26 <thinrichs> So let's all think about what talks might be good at Austin, and put some time into thinking through the coding it would take (if any).
00:30:08 <thinrichs> I'm always happy to be someone you can bounce ideas off of.
00:30:25 <thinrichs> Questions or comments?
00:31:26 <thinrichs> #topic open discussion
00:31:52 <ekcs> question on distributed arch: as I understand it we should work on things assuming the same interface currently provided by deepsix right?
00:32:45 <pballand> ekcs: it wont’ be exactly the same (the six-pack references will be gone from the names, for example) but should be similar
00:33:00 <thinrichs> pballand: does the code that's up in review have the interface?
00:33:36 <pballand> not quite - I stripped it all out to simplify the code until the messagign was working
00:34:22 <ekcs> pballand: got it. I understand it won't be exactly the same. but the plan is to work as if it will be, then make the minor changes later. As I understand it.
00:34:37 <thinrichs> ekcs: sounds like the right plan to me.
00:34:49 <ekcs> Question: Whom should I talk to if I have trouble understanding deepsix? Not much doc there. So far I'm doing fine figuring it out from the code, but I may need help later from someone who knows it well.
00:35:19 <thinrichs> ekcs: I would think you shouldn't spend much time learning deepsix.  It's about to be thrown out.
00:36:07 <thinrichs> ekcs: depending on how much time you're taking off at the holidays, you might just write the basic code assuming the new DSENode is in place and then wait to test until it lands.
00:36:08 <ekcs> thinrichs: Right. I guess I mean understanding how to use deepsix (interface) and not so much how deepsix works inside.
00:36:41 <thinrichs> ekcs: The only bits we ever used are the pub/sub and rpc functionality.
00:36:53 <ekcs> thinrichs: okay.
00:37:04 <thinrichs> pub/sub for datasources sending data on the bus, and rpc for executing actions on datasource drivers
00:37:38 <thinrichs> I did the original integration of the datasources, the policy engine, and deepsix, so I'm happy to help.
00:37:54 <ekcs> thinrichs: okay thanks.
00:37:59 <thinrichs> pballand: while writing a new Dse did you learn much about the old one?
00:38:36 <pballand> not really :-/
00:39:14 <thinrichs> pballand: just thought I'd ask
00:39:30 <thinrichs> I can't think of anyone else who would have worked much with the DSE then.
00:39:46 <ekcs> okay.
00:39:51 <thinrichs> ramineni: want to give us an update on your tempest and the gate fixes?  Anything new or interesting going on there?
00:40:10 <ramineni> thinrichs: ok
00:41:05 <ramineni> thinrichs: i got it wrkng the final pieces left for tempest tests ( basic ops and HA ), they are up for review
00:41:08 <ramineni> :)
00:41:37 <thinrichs> Great!
00:41:45 <thinrichs> Those have been broken a long while.
00:41:45 <ramineni> thinrichs: but we need a fix in tempest also for them to wrk , so marked as WIP now, until that patch is merged
00:42:04 <thinrichs> It's good we're getting them back in place.
00:42:20 <ramineni> thinrichs: ya
00:42:56 <thinrichs> I saw another update for projects.yaml.  That's to get tests working on stable/kilo and stable/liberty, right?
00:43:13 <ramineni> thinrichs: i think gate is green now , but might be broken again , as tempest have 3 patches about to merge which cause our gate to break again
00:43:32 <ramineni> thinrichs: yes , but no reviews on that :(
00:44:09 <ramineni> thinrichs: tried to ping infra on reviewing them , but i guess they are busy
00:44:48 <thinrichs> I'll try to review it quickly.  Maybe that'll bump it up on the stack.  Though it's close to the holidays, so that could be the  problem.
00:45:05 <ramineni> thinrichs: just added to run tempest like before without plugin for stable branches , which i forgot to add before
00:45:19 <ramineni> thinrichs: that would be great , thanks
00:46:40 <thinrichs> ramineni: done
00:46:54 <ramineni> thinrichs: thanks
00:47:03 <thinrichs> That's it from me.  Does anyone have anything else?  Or should we end a little early?
00:47:58 <masahito> no meeting in next week, right?
00:48:12 <thinrichs> No meeting next week.
00:48:21 <thinrichs> masahito: thanks for reminding me
00:48:54 <thinrichs> We'll meet again the week of Jan 3-9
00:49:05 <thinrichs> masahito: did you have a quick status update for us?
00:49:54 <masahito> ok
00:50:22 <masahito> I'm thinking how we add TRANSLATORS for push-driver.
00:51:14 <masahito> Existing services, like Nova, should use its TRANSLATORS
00:52:10 <thinrichs> One thought: what if we had the Nova driver implement methods for both the push and pull version?
00:52:16 <thinrichs> So the Nova_driver would implement...
00:52:31 <thinrichs> update_from_datasource()
00:52:35 <masahito> but for others I'm wondering how to add it because writing one driver is much costs
00:52:36 <thinrichs> get_schema()
00:52:53 <thinrichs> … <whatever else is in the interface for today's pull-based drivers>
00:53:05 <thinrichs> and then also have Nova driver implement the new push-based interface:
00:53:19 <thinrichs> push(jsondata)
00:53:44 <thinrichs> Then Nova can use 1 set of translators and implement either or both of the push and pull interfaces.
00:54:13 <thinrichs> It could inherit from both the Push and Pull driver base classes.
00:54:17 <thinrichs> Is that crazy?
00:54:24 <masahito> thinrichs: sounds fine. I'll try it.
00:55:15 <thinrichs> masahito: your first draft of the push driver was great progress!
00:55:39 <thinrichs> masahito: thanks for volunteering for that feature
00:55:42 <masahito> thinrichs: thanks.
00:56:12 <thinrichs> 5 minutes left.
00:56:55 <thinrichs> Happy holidays everyone!
00:57:39 <thinrichs> #endmeeting