17:00:45 #startmeeting CongressTeamMeeting 17:00:46 Meeting started Tue Mar 31 17:00:45 2015 UTC and is due to finish in 60 minutes. The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:47 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:50 The meeting name has been set to 'congressteammeeting' 17:01:20 Let's give people a couple of minutes before we get started. 17:01:33 Hi 17:01:56 arosen1: hi 17:02:10 hi all 17:02:37 That's enough folks to get started with status updates. 17:02:39 Hi all 17:02:40 #topic status 17:02:45 bryan_att: hi! 17:02:56 arosen1: want to give your status report? 17:03:00 sure 17:03:39 so the CI should be back up and running correctly soon. Once this patch mergers in tempest we'll be good again: https://review.openstack.org/#/c/169115/4 17:04:01 lurking 17:04:03 I'm also working on a trigger api call that can be given to congress to on-demand update a specific datasource. 17:04:30 that's all i have to report for now. 17:04:48 iben: hi! 17:05:15 arosen1: that's cool stuff. We've talked about having those triggers for a long time. Good to see it's finally happening. 17:05:33 alexsyip: want to go next? 17:05:42 I have the datasource configuration synchronization working. 17:06:13 I’ve been working on a tempest test to verify it but I’ve been having to setup authentication via keystone and I’m not sure how that needs to work yet. 17:06:50 I will talk with arosen1 today to get some help. 17:07:08 that’s all. 17:07:26 So this is our first foray into high-availability. 17:07:33 Cool! 17:08:20 bryan_att: do you have any further questions, or are you finding everything you need? 17:08:37 Pierre from vmware was looking into an ansible installer for Congress - we are waiting for that to proceed with Saltstack install 17:09:13 pettori: you there? How's the ansible installer going? 17:09:21 hi guys 17:09:41 sorry the ansible installer is mostly done 17:09:46 link? 17:09:53 probably will be finish today or tomorrow 17:10:04 I will see if we can push it today and will send the link 17:10:27 it’s okay (better) to be a work in progress. 17:10:37 would never really be “done” would it? 17:10:40 sounds good 17:10:47 iben: yep 17:10:51 thanks for the update 17:11:21 other status: I'm continuing to test out Congress using devstack and further developing the use cases and requirements under the OPNFV Copper project. I'll next be writing test cases to exercise the use cases given the current Congress datasources. 17:11:44 bryan_att: that's great! We love having more use cases. 17:11:48 bryan_att: awesome! 17:12:00 My goal is to have something to demo at one or more up the upcoming conferences - NFV World, OpenStack, ONS, etc 17:12:34 This time I will have a real running platform in our labs to use - based upon OPNFV if I can get Congress installed on top of it 17:12:54 bryan_att: let us kow when you hit snags — we're happy to help. 17:13:30 sure -thanks. For now, it doesn't look like we will have any blueprints for L release though. 17:13:57 I put a couple placeholders but my investigation showed that the Congress code has a pretty comprehensive set of data exposed 17:14:06 So I will start there and test out the use cases 17:14:15 Makes sense. 17:14:39 the pub-sub framework for realtime event notification is also something I am interested in 17:14:57 The reason we don't have blueprints for Liberty is that there's been a bunch of activity recently that we're trying to help along. 17:15:13 And we have some code to write for a demo for one of the OpenStack Congress talks. 17:15:40 Right now I'm planning for a feature freeze in a little more than a week (with the exception of code needed for the demo). 17:15:59 For future releases, has there been any discussion about real-time policy-related data notifications from openstack components? 17:16:00 Then we'll have a month to debug and finalize a release candidate. 17:16:43 bryan_att: the only thing we've considered is arosen1's work on triggers to tell us when to refresh our data cache. 17:16:58 OK, I will watch that 17:17:23 Doing a real-time data pull each time we answer a query would be expensive. 17:17:36 in general. 17:17:59 But having the ability to pick and choose how fresh the data needs to be, with one option being real-time fresh would be nice. 17:18:18 yes, I meant reall setting a trigger for an openstack component to tell Congress when a certain data change occurred. Similar to some of the things proposed for Cielometer. 17:18:35 bryan_att: yep—that's what arosen1 is working on right now. 17:18:40 cool 17:19:06 jwy: how's your stuff going? 17:19:45 good, i have patches for create policy, delete policy, and delete policy rule: https://review.openstack.org/#/q/status:open+project:stackforge/congress+branch:master+topic:bp/horizon-create-policies,n,z 17:19:52 the first 3 are ready for review 17:20:13 expecting to be done with the remaining, create policy rule, this week 17:20:45 been helping Rajdeep and Yali with their blueprints 17:21:10 Cool! Can I have a couple of volunteers familiar with Horizon to review? 17:21:54 Ok then—we'll volunteer offline. 17:22:04 jwy: don't be shy about poking people. :) 17:22:24 yep, thanks 17:22:27 that's all i have 17:22:36 jwy: thanks 17:22:56 stevenldt: I see you've been making progress this week. Want to tell us about it? 17:23:06 Sure. 17:23:35 I finish up the action execution interface. The functional is in. Last set of changes are ready for review. 17:23:50 Next thing would be some tempest tests and additional API 17:24:32 I also make changes to Murano driver to have a workaround for recursive rules used in Murano 17:25:00 stevenldt: good stuff! Action execution is important for the reactive form of enforcement we're hoping to get into kilo. 17:25:18 More on our progress on reactive enforcement in a bit. 17:25:37 That's about it for me. 17:25:48 stevenldt: thanks! 17:25:53 One more thought on this topic... 17:26:10 for those that don't know, our policy engine doesn't currently support recursion. 17:26:19 (There's an experimental version that does, but it's not ready yet.) 17:26:55 The integration with Murano needs recursion, so instead of putting it into the policy itself, stevenldt put that logic into the datasource driver. 17:27:14 We'll need to experiment whether this is the right approach long-term, but for now it seems to be the right way to go. 17:27:58 I have a few items to report on, but before that... 17:28:06 did I miss anyone who has a status report for us? 17:29:18 A couple of things to report on then. 17:30:07 First the Congress project is applying to become part of OpenStack proper (what used to be known as 'incubated') 17:30:28 #link https://review.openstack.org/#/c/165204/ 17:30:51 I'll give everyone a minute to look that over. 17:31:03 Of particular interest are the comments from the community at the bottom. 17:33:55 Given our plans to do a release on roughly the usual schedule for kilo, there don't seem to be any major objections. 17:34:24 But we should do what we can to make the official kilo release date. 17:34:32 sounds good to me. 17:34:38 I think we should definitely be able to do that. 17:34:44 seems a go from the comments 17:35:26 exciting 17:35:34 The only potential holdup for hitting the release is the explicit-reactive-enforcement work. 17:35:43 #link https://blueprints.launchpad.net/congress/+spec/explicit-reactive-enforcement 17:36:46 We've seen a bunch of work being done on it, and it was one of the main features we planned for kilo. 17:36:56 And I think we're pretty close to getting it done. 17:37:12 If you scroll to the bottom of that blueprint, you'll see it's really 4 blueprints that are needed for completion. 17:37:37 The 3 dependent blueprints are nearing completion now. 17:38:25 And something close to the 4th one (an application of the policy-engine triggers) was just pushed for review this week. 17:38:49 https://review.openstack.org/#/c/134340/ 17:39:02 So in short, we're getting close. 17:39:18 With some luck we can get it all in place by the end of next week. 17:39:37 So that's why it seems like the right time for feature freeze and to cut a kilo3 milestone release. 17:40:19 That'll give us a 3+ weeks to bug-fix, sync docs, etc. before the official kilo release. 17:40:25 Sound like a plan? 17:40:34 sounds good to me. 17:40:42 yep 17:42:53 I guess I should also mention that I've started working on a delegation demo for one of our talks at the summit. 17:42:56 Some links… 17:43:21 PoC design doc: 17:43:22 https://docs.google.com/document/d/1ksDilJYXV-5AXWON8PLMedDKr9NpS8VbT0jIy_MIEtI/edit# 17:43:35 People who signed up for some of the pieces of the work: 17:43:42 https://docs.google.com/spreadsheets/d/12_1TD-MED_EIGaoEhUE62IjrwbAmLgS9eX6gAo3-qKM/edit#gid=0 17:44:11 Skeleton for our first domain-specific policy engine driver 17:44:11 https://review.openstack.org/#/c/167407/ 17:44:26 There's plenty to do here, so let me know if you're interested. 17:44:55 I think that's it for my status report. 17:45:14 Let's open it up for discussion now. 17:45:17 #topic open discussion 17:46:12 thanks for the delegation doc links. very interesting, I will review with our policy team 17:46:32 bryan_att: let me know if you have any questions! 17:47:37 thinrichs: i saw on the openstack-dev list that there's a meeting today for PTLs and cross-project liaisons to discuss design summit content. will you be attending? 17:48:16 do you know if sarob got any info last week about getting a congress time slot during the design summit? 17:48:41 jwy: when I talked to him Friday he hadn't heard back about the design session yet. 17:49:06 We should probably ping him about it later in the week. 17:49:37 jwy: I hadn't heard about the PTLs and the design summit. Do you have a link handy? 17:49:58 #link https://wiki.openstack.org/wiki/Meetings/CrossProjectMeeting 17:52:23 I'll definitely try to attend that—or at least lurk. 17:52:27 jwy: thanks for the pointer! 17:52:35 thinrichs: cool, np 17:53:42 Anything else? 17:54:26 OK. Let's call it for today then. 17:54:30 Thanks all! 17:54:39 #endmeeting