12:59:52 <Qiming> #startmeeting senlin
12:59:53 <openstack> Meeting started Tue Nov  3 12:59:52 2015 UTC and is due to finish in 60 minutes.  The chair is Qiming. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:59:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:59:56 <openstack> The meeting name has been set to 'senlin'
13:00:03 <Qiming> hello
13:00:06 <yanyanhu> hi
13:00:08 <elynn> Hi
13:00:15 <Qiming> evening
13:00:23 <HAIWEI> hi
13:00:35 <Qiming> #link https://wiki.openstack.org/wiki/Meetings/SenlinAgenda
13:00:47 <Qiming> please check the meeting agenda and see if you have things to add
13:01:02 <Qiming> the only thing I have in mind is about mitaka plan
13:01:30 <Qiming> let's start with it
13:01:48 <Qiming> #topic Mitaka work items
13:02:21 <Qiming> so we have several places that document our previous work items/plan
13:02:55 <Qiming> yanyanhu has helped revise the TODO.rst file
13:03:04 <yanyanhu> yes, we now have etherpad, TODO.rst and also blueprint I think
13:03:19 <yanyanhu> https://review.openstack.org/240764
13:03:21 <Qiming> https://review.openstack.org/#/c/240764/2/TODO.rst
13:03:59 <Qiming> we may need to consolidate backlogs into the TODO.rst file
13:04:19 <Qiming> or at most we can maintain a single etherpad page for collaborative editing
13:04:46 <yanyanhu> agree with this
13:05:19 <Qiming> let's do some editings, :)
13:05:42 <HAIWEI> just saw the Mitaka tasks, about placement policy, is Multi-region different from multi-cloud?
13:05:45 <Qiming> I'm moving liberty items to the mitaka items page
13:05:49 <yanyanhu> yes
13:05:56 <yanyanhu> they should be different
13:06:31 <HAIWEI> i thought they were the same
13:07:11 <Qiming> multi-cloud may have more options than multi-region, HAIWEI
13:07:41 <Qiming> liberty items cleaned, :)
13:09:18 <yanyanhu> Qiming, do we need to list all possible workitems in etherpad?
13:09:46 <Qiming> let's move our meetup items into the mitaka items, and then we compare the list with the TODO.rst revision
13:09:53 <yanyanhu> I guess maybe we can just record those items with lower priority in TODO?
13:09:58 <yanyanhu> ok
13:10:00 <HAIWEI> Or specify the work items to Mitaka-1?
13:10:15 <Qiming> we cannot edit the TODO.rst file directly
13:10:26 <elynn> Do we need a specs folder for BP design?
13:10:50 <HAIWEI> maybe we should
13:11:09 <Qiming> HAIWEI, we need a list of all work items, then we prioritize them
13:11:21 <Qiming> then we set our goal for mitaka-1
13:11:31 <HAIWEI> ok
13:12:01 <Qiming> elynn, we have been doing things a little bit different from other openstack projects
13:12:34 <Qiming> we have two files maintained in the code base: TODO.rst file for workitems that doesn't require a blueprint or spec
13:13:20 <Qiming> FEATURES.rst for things (most of which are about mid-term/long-term goals) we will do in future, and they all needs some detailed design before being worked on
13:13:28 <yanyanhu> or maybe each time we pick up a work item from TODO.rst, we file a blueprint/spec for it?
13:14:05 <yanyanhu> and when the bp/spec is completed, we remove the item from TODO list
13:14:10 <Qiming> items in TODO.rst are not the place for bug reports
13:14:17 <yanyanhu> yep
13:14:25 <Qiming> please use launchpad for bug reports, we track bugs there
13:15:15 <HAIWEI> yes, I think we should be strict for the bug fix, if the patch is fixing a bug, we should file the bug in launchpad first
13:15:15 <Qiming> is this clear for everyone?
13:15:29 <yanyanhu> HAIWEI, agree
13:15:30 <Qiming> agreed
13:15:31 <yanyanhu> yes
13:15:34 <elynn> yes
13:15:40 <lixinhui> yes
13:15:45 <HAIWEI> I am clear about it
13:16:23 <Qiming> okay, so we will have several things to maintain, mostly follow the community practices
13:16:38 <Qiming> speaking of specs, we may need it soon
13:17:31 <Qiming> if it is about a new feature, or a huge refactor, we need a plan/discussion before working on it
13:17:34 <Qiming> that's what specs are used for
13:17:53 <yanyanhu> yea
13:18:36 <Qiming> my etherpad connection is very unstable at the moment
13:19:56 <Qiming> seems having two etherpad pages open is not a good idea
13:20:32 <Qiming> let's focus on the mitaka meetup page
13:20:49 <yanyanhu> ok
13:21:00 <Qiming> delete items that are already merged into the TODO.rst file
13:21:24 <Qiming> then we have a clear picture of where we are heading, :)
13:21:41 <yanyanhu> hi, Qiming, or maybe we keep this page for some people we want to check the result of our discussion in Tokyo meetup?
13:21:49 <Qiming> API consistency has been merged, right?
13:22:01 <yanyanhu> since somebody may not got the chance to join the discussion
13:22:05 <yanyanhu> Qiming, yes
13:22:10 <elynn> That's me
13:22:15 <Qiming> yanyanhu, don't think so, we'd better maintain information in the same place
13:22:27 <Qiming> scatter things everywhere will only cause confusion
13:22:37 <yanyanhu> ok
13:23:02 <Qiming> need check existing API WG guidelines and find the gaps
13:23:12 <Qiming> this is not merged, right?
13:23:40 <yanyanhu> yes, I think so
13:23:59 <Qiming> API versioning is not merged, but it is of low priority, let's focus on v1 at present
13:24:45 <Qiming> yanyanhu, for the items mentioned, I'm leaving them there, pls help merge them to TODO.rst and FEATURE.rst respectively, offline
13:24:55 <yanyanhu> ok
13:26:14 <Qiming> elynn, lixinhui ... you were not at the meetup
13:26:26 <Qiming> please speak up if you have questions, :)
13:26:42 <elynn> So finally, which etherpad do we maintain?
13:26:59 <Qiming> senlin-mitaka-workitems
13:27:19 <Qiming> on etherpad, we only maintain this single page, for weekly meetings
13:27:59 <elynn> ok, and just saw that webhook will be rename to receiver?
13:28:21 <Qiming> right, that was discussed during meetup
13:28:36 <Qiming> webhook is a special type of receiver
13:29:07 <Qiming> we want the senlin API as stable as possible, after all, it is a contract with senlin service users
13:29:35 <Qiming> other types of receivers may include message queues etc
13:29:36 <elynn> I see.
13:30:03 <Qiming> if we have webhook only api, we will have a lot of difficulty to deprecate it
13:30:22 <Qiming> the lessons we learnt ... don't even try to deprecate anything, :)
13:31:05 <elynn> That's right, backward compatible is not an easy thing.
13:31:05 <Qiming> elynn, heat resource type support is clear right?
13:31:40 <Qiming> it should be okay to say that we only support one type of receivers
13:31:57 <elynn> Yes, after the senlinclient patch is merged, I can start to work on senlin plugin in heat.
13:32:16 <Qiming> cluster, node, profile ... these resources are merged into sdk now
13:32:50 <Qiming> so no blocker to support them in heat, except for the senlinclient refactor
13:33:02 <Qiming> yanyanhu has helped tested the new interface, it seems work well
13:33:24 <elynn> But we still need get_file support for senlinclient and heatclient.
13:33:26 <yanyanhu> yes, I think it has been ready
13:33:40 <Qiming> we need to bump client version to 0.1.5 soon, so your heat resource type work and the senlin-dashboard can move on
13:34:27 <elynn> yes, needs a new version after refactor.
13:35:06 <elynn> Speaking of openstacksdk, I notice that workitem 10. Dependency on openstack-sdk (how about using openstack-client)
13:35:17 <Qiming> hope the invocations would be smooth afterwards
13:35:46 <Qiming> elynn, I spent sometime on reading osc code
13:35:56 <Qiming> it won't serve our use case
13:36:03 <Qiming> we need to stick to sdk
13:36:09 <Qiming> osc is only about command line
13:36:23 <elynn> Ok, so this is not an workitem yet. got it.
13:36:32 <Qiming> we talked to Brian in Tokyo, the code review will catch up
13:38:13 <Qiming> yanyanhu, have we documented TOSCA related items?
13:38:23 <yanyanhu> not yet, Qiming
13:38:32 <Qiming> seems to me we need a couple of specs before getting hands on them
13:39:11 <Qiming> okay, need to work with Matt on it
13:39:34 <yanyanhu> yes, I think we need some specs to clarify more detail about this job
13:40:36 <Qiming> we do no convergence, but we need to manage objects we created on behalf of users, that was the conclusion, right?
13:40:48 <yanyanhu> yes
13:41:01 <yanyanhu> that's right since we are cluster management service
13:41:05 <Qiming> policy data passing thing, merged?
13:41:06 <yanyanhu> not just cluster deployment service
13:41:20 <yanyanhu> let me check
13:42:00 <yanyanhu> hi, Qiming, not yet
13:42:10 <Qiming> ok, leaving it there
13:42:17 <yanyanhu> since I was not quite sure how to describe the workitem accurately
13:42:26 <yanyanhu> will think through about it
13:42:49 <Qiming> just try your best, we are a team, we can help polish it once there is a draft
13:43:07 <yanyanhu> ok :)
13:43:23 <Qiming> the only big mistake we should avoid is we forget what we have discussed, :)
13:43:33 <yanyanhu> right
13:43:50 <Qiming> having all these captured into the TODO.rst file and/or the FEATURE.rst file will help
13:44:20 <Qiming> anything we neeed to delete from the meetup page?
13:45:06 <yanyanhu> - Scavenger of objects (should be configurable.)
13:45:07 <yanyanhu> - object status consitency (clearly document )
13:45:31 <Qiming> those two already merged?
13:45:32 <yanyanhu> hi, Qiming, I think we can also remove these two items which are listed under Housekeeping catalog
13:45:49 <yanyanhu> yes, they were also added in today's patch
13:46:00 <Qiming> cool
13:47:04 <Qiming> okay, after the meeting, please help consolidate the rest into the two rst files
13:47:17 <yanyanhu> ok
13:47:19 <Qiming> time to move on, :)
13:47:58 <Qiming> don't think we'll have time to work on this today: https://etherpad.openstack.org/p/senlin-mitaka-workitems
13:48:18 <Qiming> this will be our focus next meeting,
13:48:32 <HAIWEI> discuss about it?
13:48:32 <yanyanhu> after we update TODO.rst, we can sync them
13:48:36 <Qiming> we review your patch and dump things here
13:49:12 <Qiming> the current list for mitaka is too long
13:49:28 <Qiming> we won't have a chance to complete all of them
13:50:00 <yanyanhu> also think so
13:50:07 <Qiming> need to prioritize things, set a clear goal for m-1
13:50:23 <HAIWEI> yes
13:50:36 <HAIWEI> when is the end of m-1
13:50:56 <yanyanhu> by the end of this year?
13:51:09 <Qiming> https://launchpad.net/senlin/+milestones
13:51:27 <Qiming> we have only one month for m-1
13:51:41 <HAIWEI> only a month away
13:51:45 <yanyanhu> ... short period
13:51:54 <Qiming> yup, got to be realistic
13:52:22 <Qiming> so the top priority in my view is about stability and heat resource type
13:52:36 <Qiming> all features can be added later on
13:52:51 <Qiming> they can be driven by requests
13:54:03 <Qiming> let's continue this next week
13:54:12 <Qiming> #topic open discussions
13:54:36 <Qiming> see if any thing guys want to bring up?
13:54:36 <yanyanhu> nice, will clean all items left in meetup page tomorrow and wait for you guys comments
13:54:43 <yanyanhu> nope from me
13:54:53 <Qiming> you already got some from me, :)
13:55:08 <yanyanhu> will check it :)
13:55:11 <HAIWEI> it seems we got some new members
13:55:42 <HAIWEI> hope more and more people will join
13:55:47 <Qiming> yes, that is getting things nicer :)
13:56:00 <HAIWEI> that's all from me
13:56:07 <yanyanhu> HAIWEI, I noticed yuanying san is also in the irc channel :)
13:56:33 <HAIWEI> yes, but I think he will not contribute :)
13:56:46 <HAIWEI> he mainly focuses on magnum
13:56:52 <Qiming> buy hime some sake
13:56:55 <yanyanhu> np, just stand there to support us :)
13:57:24 <Qiming> HAIWEI, we need to continue client test cases
13:57:26 <HAIWEI> anyway, we want to use senlin for our local team
13:57:38 <Qiming> at least for the shell module
13:57:38 <HAIWEI> ok, Qiming
13:58:12 <elynn> I have some concern about the patch for senlinclient.
13:58:25 <elynn> about profile_create
13:58:30 <Qiming> the client module will change, when we have better support from sdk
13:58:36 <elynn> I will comment in that patch.
13:58:37 <HAIWEI> you start a patch, and we cooperate with the job, Qiming
13:58:58 <Qiming> okay, thanks, elynn
13:59:44 <Qiming> HAIWEI, I think I have already done that
13:59:44 <Qiming> maybe need more, ;)
13:59:44 <Qiming> let's release this channel, time's up
13:59:44 <Qiming> #endmeeting