16:00:20 <rakhmerov> #startmeeting mistral
16:00:21 <openstack> Meeting started Mon Jun 16 16:00:20 2014 UTC and is due to finish in 60 minutes.  The chair is rakhmerov. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:24 <openstack> The meeting name has been set to 'mistral'
16:00:30 <rakhmerov> hi
16:00:59 <m4dcoder> hi
16:01:07 <rakhmerov> how are you?
16:01:17 <m4dcoder> doing good.  how about yourself?
16:01:28 <rakhmerov> still alive :)
16:01:34 <dzimine> hi all
16:01:39 <rakhmerov> hi Dmitri
16:02:02 <rakhmerov> so let's begin
16:02:54 <dzimine> renat welcome back :)
16:02:58 <rakhmerov> thanks
16:03:00 <rakhmerov> can we start with quick member-by-member update today?
16:03:32 <rakhmerov> I read the previous meeting log so am a little bit aware of what you guys were doing
16:03:50 <dzimine> I’ll start as there’s little to update: didn’t have time to continue ‘start-task’ blueprint last week unfortunately, so only did reviews
16:04:11 <rakhmerov> ok, that's fine
16:04:34 <rakhmerov> dzimine, do you see any potential problems with this task?
16:04:51 <dzimine> no. it’s just a bunch of updates all over the place
16:04:58 <rakhmerov> one thing I'm worried about is that we must be really careful about changing our API
16:05:06 <dzimine> I already have the draft impl.
16:05:20 <dzimine> let’s talk about API later today. this is a key topic.
16:05:21 <rakhmerov> Angus warned us many times that we should be carefully versioning our API
16:05:39 <rakhmerov> ok, sure, let's postpone it
16:05:50 <rakhmerov> m4dcoder?
16:05:53 <m4dcoder> proposed design for https://blueprints.launchpad.net/mistral/+spec/mistral-engine-executor-protocol on ML. also had some questions on parallelism referenced in the bp.  waiting for responses.  i have some other priority work to do this week so probably won't pick up this bp work until next week.
16:06:41 <rakhmerov> I see
16:06:51 <rakhmerov> could you please find a link to this thread?
16:07:06 <rakhmerov> there's 1500 emails in my inbox and I didn't process them yet )
16:07:07 <m4dcoder> http://lists.openstack.org/pipermail/openstack-dev/2014-June/037547.html
16:07:17 <rakhmerov> thanks a lot
16:07:20 <m4dcoder> np
16:07:46 <rakhmerov> #info Winson's desing for engine-executor protocol: http://lists.openstack.org/pipermail/openstack-dev/2014-June/037547.html
16:08:08 <rakhmerov> I'll look at it tomorrow morning and reply to you
16:08:17 <m4dcoder> thanks!
16:08:18 <rakhmerov> that's pretty important for us
16:09:15 <rakhmerov> Nikolay, are you with us today?
16:09:51 <rakhmerov> he's probably not
16:10:10 <rakhmerov> Timur also said he couldn't make it to be at the meeting
16:10:22 <rakhmerov> so, as far as myself
16:10:53 <dzimine> is Angus around?
16:11:29 <rakhmerov> I spent most of the time on reading emails, BP comments and talking to folks to find out all the news
16:11:56 <rakhmerov> and I also started thinking on the big engine refactoring that we agreed on at the summit
16:12:21 <rakhmerov> have some ideas written down but nothing is really shaped out yet
16:13:21 <rakhmerov> my plan is to spend some time researching ansible and other wf technologies first since I feel I have lots of gaps in my awareness
16:13:47 <rakhmerov> I also reviewed several patches today and sent a couple of my own patches
16:14:35 <rakhmerov> btw, dzimine, I agree with you suggestion to hold on making our devstack gate voting
16:15:15 <rakhmerov> there's a patch already but it'll be hanging for a while with -1 till we see it's comfortable to activate it
16:15:41 <rakhmerov> so anything else for now?
16:15:58 <rakhmerov> I would switch to open discussion and talk about API
16:16:14 <rakhmerov> unless we have anything else: roadblocks, questions etc.
16:16:31 <dzimine> API is the #1.
16:16:37 <rakhmerov> ok
16:16:42 <rakhmerov> #topic Open Discussion
16:16:52 <rakhmerov> ok, fire away Dmitri
16:17:07 <rakhmerov> what are your thoughts?
16:17:19 <dzimine> Just like u, I spoke with Angus: they’re ready to use Mistral and once they do we can’t break things.
16:17:28 <dzimine> But, we are making breaking changes on API.
16:17:29 <rakhmerov> yes
16:17:36 <rakhmerov> which ones?
16:18:01 <dzimine> start-task, workflow-in-workflow, and exposing worklfows, triggeres, actiosn as 1st class entities.
16:18:20 <rakhmerov> oooh, but it's all in progress now
16:18:24 <rakhmerov> which is good
16:18:31 <dzimine> So, how much of these can we do before we freeze the API?
16:18:36 <rakhmerov> I believe so far we didn't break anything
16:19:07 <rakhmerov> I would limit those changes in v1 by start-task
16:19:26 <rakhmerov> other things should be done as a part of the next API version
16:19:52 <dzimine> What is the deprecation story?
16:19:54 <rakhmerov> because seemingly start-task thing is almost backwards compatible, am I right?
16:20:25 <rakhmerov> I mean "task" parameter could be just ommitted
16:20:32 <rakhmerov> deprecation?
16:20:42 <rakhmerov> you mean "requires"?
16:22:11 <dzimine> yes ‘task’ can be omitted and ignored in API call. Must present in DSL, though.
16:22:28 <rakhmerov> if you're talking about "requires" I'm ready to deprecate it now. However, we have to run it by our current stackholders: Fuel and Solum
16:22:50 <rakhmerov> yeah, DSL will slightly change
16:23:20 <rakhmerov> which is bad but I think we can afford this change now
16:23:51 <dzimine> Here’s a suggestion: include start task and deprecate require’ in v1, and do the rest in v2.
16:24:07 <rakhmerov> agree
16:24:27 <dzimine> An alternative would be to ask Angus to wait and help with the API (which he offered) and do the API cleanup in V1.
16:24:34 <rakhmerov> but again, let's make sure we won't steal a goodie from FUEL and SOLUM
16:24:40 <rakhmerov> I mean "requies"
16:25:17 <rakhmerov> that's a good alternative
16:25:25 <rakhmerov> let me talk to him tomorrow
16:25:34 <dzimine> I wonder if they can wait.
16:25:43 <rakhmerov> I just need to sync up with him and see what he's up to
16:25:49 <rakhmerov> yes, dunno
16:26:03 <dzimine> with 2 versions of API we’ll need to support 2 sets of tests, etc, it’s too early. If they only can wait for a bit.
16:26:09 <rakhmerov> anyway, I wouldn't include other changes in V1 like multiple workflows etc.
16:26:21 <rakhmerov> only start-task and "requires" deprecation
16:26:45 <rakhmerov> yeah, good point about tests and whatever else
16:27:23 <rakhmerov> ok, so I think we have an agreement on this
16:27:53 <rakhmerov> the only thing I would suggest is that we shouldn't hurry with this
16:28:11 <rakhmerov> we gotta be really really careful and think twice
16:28:32 <rakhmerov> and discuss things in ML
16:29:03 <rakhmerov> m4dcoder, just recalled I had a question for you
16:29:12 <rakhmerov> it's minor but I'm just curious
16:29:24 <m4dcoder> shoot
16:29:27 <rakhmerov> why did you rename mistral.conf.example to mistral.conf.sample?
16:29:45 <m4dcoder> generate_sample.sh generates sample file with this name.
16:29:49 <rakhmerov> was there any strong reason or it was just a preference?
16:29:57 <rakhmerov> ooh
16:30:00 <m4dcoder> also if we use check_todate.sh, it's looking for that file.
16:30:24 <m4dcoder> more consistent then preference.
16:30:25 <rakhmerov> ok, so it's sort of a standard naming in OpenStack, right?
16:30:28 <m4dcoder> yep
16:30:33 <rakhmerov> ok, cool
16:31:02 <rakhmerov> I'm just asking because something got broken because of that but that's fine, we fixed it
16:31:23 <m4dcoder> cool
16:31:24 <rakhmerov> and I also renamed other example files in this manner too
16:31:30 <rakhmerov> ok
16:32:05 <rakhmerov> dzimine, a question to you
16:32:11 <dzimine> fire
16:32:35 <rakhmerov> did you already assigned all the BPs to your folks as you planned about 3 weeks ago?
16:32:59 <rakhmerov> so that I know who is supposed to do what
16:33:24 <rakhmerov> in terms of amount of time you already told me so I just would like to know some detailes
16:33:25 <dzimine> no: I am supposed to do this but I only do once they’re available to pick it up.
16:33:44 <rakhmerov> ok
16:34:09 <rakhmerov> as far as I understand, Manas is not available for now right?
16:34:44 <rakhmerov> and Kirill very little
16:34:57 <dzimine> no for Manas, yes for Kirill.
16:35:16 <rakhmerov> alright
16:36:12 <rakhmerov> btw, I had a discussion with the folks today and told again that we should be more open to you
16:37:26 <dzimine> yes, thanks: and not just to us but to the community.
16:38:28 <rakhmerov> yes, right
16:39:08 <dzimine> did you see the discussion on fuctiontests?
16:39:52 <rakhmerov> nope
16:39:55 <rakhmerov> in ML?
16:40:00 <dzimine> etherpad.
16:40:20 <dzimine> ML has a link, https://etherpad.openstack.org/p/MistralNewTestsDesign
16:40:30 <rakhmerov> can you include the link pls?
16:40:32 <rakhmerov> ok
16:41:43 <rakhmerov> is there anything that's worth some special attention or it's just FYI?
16:42:06 <dzimine> The attempt is to make as much of functional tests EASILY runnable by developers, the smaller subset which require OpenStack can be run with devstack. Once we are there, we can turn the devstack gate on.
16:42:25 <rakhmerov> I see
16:42:45 <dzimine> I want to avoid the situation when there are some tests whic you can run on dev box (or takes too much effort to run), and they fail the gate.
16:43:03 <rakhmerov> well, assummingly we should have two test sets: OpenStack related and OpenStack free
16:43:07 <dzimine> And, if we have them easy to run, we can chage the tests with the code.
16:43:47 <rakhmerov> sure, I totally agree with this
16:43:53 <dzimine> Yes that’s the proposal. Now it’s not separated yet: all API tests assumed OpenStack related, but most API calls have little to do with OpenStack.
16:43:57 <rakhmerov> it would make the process more agile
16:44:17 <rakhmerov> yes
16:44:24 <dzimine> So we all are in general agreement, we only need to give this the right priority.
16:44:32 <dzimine> let’s check with Timur.
16:44:47 <rakhmerov> ok
16:45:15 <rakhmerov> btw, we need to start getting Sergey Murashov more and more involved into this
16:45:39 <rakhmerov> he's going to replace Timur almost 100%
16:46:11 <rakhmerov> alright
16:46:23 <rakhmerov> anything else for today?
16:47:21 <rakhmerov> then let's finish the meeting
16:47:29 <rakhmerov> thanks guys for joining today
16:47:30 <dzimine> ОК
16:47:36 <dzimine> cheers.
16:47:48 <m4dcoder> bye
16:47:49 <rakhmerov> I'll see you (virtually) soon
16:47:52 <rakhmerov> bye
16:47:56 <rakhmerov> #endmeeting