17:01:45 #startmeeting murano 17:01:46 Meeting started Tue Aug 19 17:01:45 2014 UTC and is due to finish in 60 minutes. The chair is ruhe. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:47 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:50 The meeting name has been set to 'murano' 17:01:51 o/ 17:02:28 o/ 17:02:43 Can we add to agenda two items? 17:02:52 #link https://wiki.openstack.org/wiki/Meetings/MuranoAgenda#Agenda 17:02:55 gokrokve_: sure 17:02:56 1) Discuss Murano participation in programms 17:03:18 2) Discuss what we miss to fully support image based applications 17:03:48 gokrokve_: ack. these topics will go right after the action items review 17:04:05 #topic action items review 17:04:19 there was only one action item - ruhe to release python-muranoclient 0.5.4 17:04:25 and i've released it last week 17:04:40 yay! 17:05:02 just to make sure everybody's aware. we have a separate launchpad project for python-muranoclient 17:05:12 o/ 17:05:24 #link https://launchpad.net/python-muranoclient 17:05:47 so, please file client related bugs and blueprints in this project 17:06:03 ok 17:06:14 we need this because clients have their own versioning schedule 17:06:27 ok. moving to next topic 17:06:34 suggested by gokrokve_ 17:06:41 #topic Discuss Murano participation in programms 17:06:46 gokrokve_: your turn 17:06:52 https://docs.google.com/a/mirantis.com/document/d/1kYvERC49bJ_qkHL2-RjyRh9JRnZGN1SAaqGRpogIC_I/edit#heading=h.nvnyp0onyt0 17:07:06 Here is a document with high level description of what we want to do 17:07:33 can you summarize for those of us without acces? 17:07:34 In the current situation in OPenStack the best way to proceed with any official sttaus is to join some existing program 17:07:45 Murano has some overlaps with other programs 17:07:56 sjmc7: You should have an access 17:08:01 sjmc7: http://paste.openstack.org/show/97395/ 17:08:11 thanks 17:08:13 sjmc7: At least I added your e-mail to list of editor 17:08:53 So, we have identified three programs where we can put Murano by parts 17:09:11 Ctalog {glance} program for Murano API and Murano repo 17:09:23 Orchestration {heat} for Murano engine 17:09:35 Dashbard {horizon} for Murano UI 17:09:56 We spoke with PTLs of Glance and Heat. They overall accept this idea 17:10:14 here is what i think about this: last week it was the only possible way to proceed. but after thursday TC meeting and latest discussions in ML, I think that we have a good chance to go with our own program 17:10:14 we still need to convince the whole teams of Glance and Heat to really approve this 17:11:03 ruhe: Discussions are good but there is no any decision after that. We don't know what TC as a tC really thinks. Right now we heard different opinions of various TC members 17:11:55 And that is the real problem. TC can't clearly articulate what they really want and what a the game rules. Right now we are in the game where rules changed without notice 17:12:46 So the best conservative way is to show actual collaboration between teams rather then trying to blindly push for a new program 17:13:06 gokrokve_: we need to weight both options. splitting across two programs will bring a few major issues. going with our own program - is more risk 17:13:10 At least that is what I think and Jay confirms 17:13:31 Sure. that is why I brought it here 17:13:43 * going with our program is more risk in terms of our chances to be accepted 17:13:44 We need as a team understand what we want and what are the options 17:14:02 ruhe: I would sya there chances close to zero 17:14:07 the fact taht we're split across two areas is part of the problem, i think 17:14:23 TC really coniders to reduce number of programs because governance problems 17:14:35 if it was just extra glance functionality and UI, then it'd be an easier sell 17:15:08 sjmc7: Sure. That the root of th problem for any high level service. It has to do more then one thing. 17:15:30 We can add API and Artifacts to Glance - thats for sure. 17:15:39 yes. so does openstack even WANT high level services in openstack? 17:15:46 But we have engine which does not belong to Glance at all 17:16:19 sjmc7: i've read all the conversations on this topic. and i don't know. i don't think there is a general agreement 17:16:35 sjmc7: TC is biased about that. They see that this really adds value but they struggle to understand how it fits to current OpenSTack model simle service for one task. 17:16:40 i think there IS general agreement, which is: Yes, but not in my project 17:17:18 if openstackk is aimed at being a set of services to use, maybe it makes sense not to incubate higher level services 17:17:26 sjmc7: True, but we did a good job talking with other teams. Glance is quite accepting this idea of having Murano API under their program. 17:18:14 sjmc7: There are debates what is incubation. We do not apply for incubation right now. What we are looking for is to join some program and put our code under their umbrella. 17:18:36 Its not an incubation. It just official statement that we want OpenStack governance 17:18:41 gokrokve_: I can hardly see how API for managing environments is drastically different from engine so that it can be in Glance while engine cannot 17:18:53 gokrokve_: this brings another concern i have about this idea - API and engine are not usable on their own. they're supposed to work *together* 17:19:07 And the change will be done in OpenStack governance document 17:19:15 yeah - to do this, the API would just need to be using heat (+mistral 17:19:20 +whetever) on its own 17:19:22 ruhe: They can work together and still be in different programs 17:19:35 ike we use Heat and we can't live without it 17:19:50 the api and engine is an implementation detail 17:19:58 sjmc7: +1 17:20:18 what you're saying is splitting te artifact repo stuff from the envioronment management stuff 17:20:26 NOT splitting the api and engine 17:20:32 Murano engine is still valuable as it adds some additional paradigms to Heat Orchestration 17:20:47 It can help to solve problems where imperative part is crutial 17:21:08 Lice DR deployments and cross clouds deployments with sofisticated error handling 17:21:23 gokrokve_: with a price of switching to MuranoPL. So this is not exactly a help for Heat users 17:21:30 yeah. but those parts are the hardest to sell 17:22:27 stanlagun: We can find out integration path together with Heat team 17:22:43 That is why we want to be a part of program 17:22:57 to be able to talk and make a decisions together as a team 17:23:01 maybe we should brainstorm pros and cons right here. or take it offline to an etherpad document? 17:23:11 Sure. Lets do this. 17:23:28 I'd like to see Zane using MuranoPL :) 17:23:30 I just want to make sure that everybody knows what is going on. 17:23:36 offline would be good. i hope to get more guidance from our needs later today or tomrrow 17:24:02 gokrokve_: yeah. thanks for driving this. we get a lot of pressure on incubation dates 17:24:15 i'd also like like to take it offline 17:24:18 sjmc7: Cool. And please involve other HP guys like Monty to understand how it could be done and what are their perspective on this 17:24:37 Again, there is no incubation 17:24:50 It is just announcement of joining the program. 17:24:56 i have very little access to our TC 17:25:07 gokrokve_: but that's a path we take to become incubated some day. right? 17:25:09 Like TripleO which is a part of program but never incubated as a project 17:25:20 ruhe: Yes. It will be incubated 17:25:52 So we need to understand 17:26:01 a) how we feel about splitting 17:26:25 b) how to present this to TC and explain the way how we collaborate with other projects 17:26:36 c) explain this to whole community 17:26:42 it would be bithing to incubate if we splitted 17:27:01 This is a first precedent when project joins program 17:27:16 katyafervent2: That a good question and probably it belongs to tC 17:27:27 honestly, i'd prefer to go with our own program to avoid the complexity of splitting the project and organisational issues being a part of 2 programs. but if there is NO WAY to have our own program, then that's the only option 17:27:28 I don't know and we need to seek a TC guidance here 17:27:47 ruhe: It is an option which we need to explore 17:27:56 ruhe: and discuss with TC 17:28:21 Ok 17:28:28 Lets move on to the next item 17:28:28 gokrokve_: will you please create an etherpad and send it across the team to continue thinking about pros and cons offline? 17:28:33 ruhe: yes 17:28:40 gokrokve_: thanks 17:28:54 I want to make sure that we fully support image based applications 17:29:10 what do you mean by that? because i thnk we're planning to deploy that way 17:29:37 https://blueprints.launchpad.net/murano/+spec/docker-registry-in-murano 17:29:43 #action gokrokve_ to create an etherpad for team to fill pros and cons of becoming a part of two programs vs having our own program 17:29:56 sjmc7: We need to support all kinds of deployment including image based 17:30:04 don't we already support that? 17:30:07 for both Glance images with apps and containers 17:30:11 #topic support image based deployment 17:30:20 sjmc7: We do support something 17:30:34 But, I want to be 100% sure that we did not anything critical 17:30:51 The first this I see is an ability to filter proper images in UI 17:31:03 Right now I can't select Ubuntu only images 17:31:07 Or Oracle 17:31:13 Or my custom App image 17:31:14 yeah, the image fliterig def needs work. but maybe not much more complicated than now, just more flexible 17:31:23 you _can_ do that 17:31:31 e.g. we just tag via the glance api rather than the UI 17:31:37 there's a BP about this already 17:31:50 sjmc7: Sure. Just add some filtering to make sure that Murano App can find its own image probably without User interfierence 17:32:13 sjmc7: Cool. Is it scheduled for Juno 17:32:28 sjmc7: Do you need it to be implemented in Juno? 17:33:08 Everythin we do for Glance images will work for Docker too 17:33:25 i have to pay attention to a meeting here now, sorry 17:33:40 So it is better to think about it now 17:33:41 Ok 17:34:45 ruhe: can you please create a dicument summarizing waht we have for Images support in Murano 17:35:01 And some image based app example will be good 17:36:20 gokrokve_: i'm not sure about the document. i'll need your input on what items you need to be described. image based examples are on our roadmap for juno 17:37:05 ruhe: Cool. We can talk offline. Some etherpad with use cases and input from Steve 17:37:25 I want to make sure that HP has everything to start using Murano 17:37:49 Also, as we have some time 17:38:05 I would like to announce what we have in PoC stage 17:38:23 1) CloudFoundry ServiceBroker API PoC 17:38:24 gokrokve_: oh. i believe sjmc7 has it under his control 17:39:09 2) Multiple Heat Stacks deployment PoC - when application can have multiple Heat Stacks rather then one 17:39:59 3) Multicloud deployments PoC - when Murano can use user credentials to authenticate rather then token only. It will help application to deploy DR topology on different datacenters or even clouds 17:40:07 gokrokve_: I thought that number two we support out of the box, or you are talking about just some example that can be shared? 17:40:29 #topic open discussion 17:40:30 serg_melikyan: #2 is possible to support but not out of the box 17:40:50 right now stack is created in environment by default 17:41:00 and it bound tightly to it 17:41:09 gokrokve_: you run ahead of topics. please let me do my bureaucracy :) 17:41:10 So to do multiple stacks is hard 17:41:26 ruhe: sure 17:41:45 we're ok now 17:41:51 So I need volunteers to do these PoC 17:42:02 #1 is almost done to something demoable 17:42:33 #2 - some discussions with Stan and some initial implementations, but we need to rewrite instance class and the way we do networking 17:42:57 #3 is not even designed, but really important and requested by customers 17:43:33 I hope we will not need to assign volunteers. 17:43:57 gokrokve_: i believe we can discuss that internally. that's not an appropriate place 17:44:21 Probably guys from other companies what to do this 17:44:27 It is not Mirantis specific 17:45:16 So lets keep this list of PoC somewhere and anybody can actually work on them 17:45:45 As our current policy is to approve BPs with some implementaiton details we can't really work without PoCs 17:45:47 HP folks have a lot on their own at this moment. I'm not sure anyone of newcomers would be able to tackle such complex scenarios 17:46:13 ruhe: True, but it does not mean that we need to keep this internal 17:46:31 gokrokve_: yeah. it's always good to have a list of future features and prioritise it taking multiple inputs into account 17:46:35 I want to make sure that averybody understands where we are going 17:47:11 sure 17:47:24 anything else to discuss today? 17:47:40 Nope, not from my side 17:47:52 Let me go from the stage :-) 17:48:11 ok 17:49:43 just an update from me - migration to oslo.db and other incubated oslo projects is going well. i don't want to leave murano with deprecated unsupport code from oslo-incubator while the rest of community is migrating to those new libs 17:49:56 *unsupported 17:50:38 i want to ask about blueprint 17:50:45 sure 17:51:47 ruhe could you provide a link please) 17:52:02 i cant copy 17:52:04 katyafervent2: https://blueprints.launchpad.net/murano/+spec/dynamic-ui-specify-no-explicit-name-field this one? 17:52:18 yes 17:52:33 did we finish with discussion? 17:52:41 katyafervent2: i don't think so 17:52:48 or should we continue to discuss 17:53:23 so when should we continue? 17:53:23 i don't see +1/+2 comments. from those who participated in the discussion 17:54:04 we had a debate about this, but didn't reach a consensus 17:54:06 everyone who is argeed on this blueprint please leave +2 comments 17:54:44 katyafervent2: you have half of the team in the same office with you :) 17:54:55 and the second half is always in IRC 17:55:55 ok, so i need to get positive comments 17:56:35 yes, that's our policy now. we need generall agreement on the blueprint before it gets an approval 17:57:10 katyafervent2: ping me if you don't get responses. i'll help you to reach out the core team 17:57:18 do we have anything else to discuss today? 17:58:21 ok. thanks everyone 17:58:29 #endmeeting