13:02:00 <ruijie_> #startmeeting senlin
13:02:01 <openstack> Meeting started Tue Sep 19 13:02:00 2017 UTC and is due to finish in 60 minutes.  The chair is ruijie_. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:02:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:02:05 <openstack> The meeting name has been set to 'senlin'
13:02:05 <ruijie_> here we go ..
13:02:10 <Qiming> :)
13:02:12 <ruijie_> sorry
13:02:20 <ruijie_> https://wiki.openstack.org/wiki/Meetings/SenlinAgenda
13:02:24 <longkb_> Hi everyone! It nice to meet you guys
13:02:32 <ruijie_> please feel free to add items to the agenda
13:02:36 <ruijie_> hi longkb_
13:03:32 <ruijie_> let's wait for a while :)
13:03:47 <ruijie_> hi chenyb4_
13:04:48 <ruijie_> okay, let's get started
13:04:58 <ruijie_> https://etherpad.openstack.org/p/senlin-queens-workitems
13:05:42 <ruijie_> hi Qiming, we discussed about the extra field/policy to add hook to scaling process
13:06:15 <Qiming> yes
13:07:40 <ruijie_> so, that policy could contains how to get approved or how to send notifications?
13:08:03 <Qiming> there are two sub-topics here
13:08:40 <Qiming> one is about how to inject something into an API request to indicate whether a request is an "original" one
13:09:45 <Qiming> the other one is about how to add hooks (possibly a general policy) that examines this 'something' and decide whether a hook will be triggered
13:10:18 <Qiming> we are supposed to be following the global api-wg guidelines, for api requests and responses
13:10:45 <ruijie_> by the "original" request, you mean we will send a request or notification to someone, and they will call back?
13:11:25 <Qiming> yes
13:13:00 <ruijie_> once the "original" comes in, we need to record it somewhere to wait for callback, after that the process could pass to next step?
13:15:29 <Qiming> we don't wait
13:15:59 <Qiming> that request has been "processed" and "redirected"
13:17:41 <Qiming> these kind of info are have to be either sent in the body or embedded in the header
13:18:41 <Qiming> if we carry these info in header, it would be something like this: https://developer.mozilla.org/en-US/docs/Web/HTTP/Access_control_CORS
13:18:59 <ruijie_> request(with special field) --> create action with special status --> return/call/notify ; callback to trigger that action --> scale in/out
13:20:56 <Qiming> the original request doesn't have to be changed
13:21:13 <Qiming> most of the time, requests should be processed as usual
13:21:37 <Qiming> it is only those requests mutated by a third party software that need a special tag
13:23:11 <ruijie_> okay, Qiming, may need to look forward to it :) thought it a little bit simple
13:24:06 <Qiming> to get it done in a disciplined way is not straightforward
13:24:25 <Qiming> although get it done in an adhoc way is simple enough
13:25:14 <Qiming> reference: http://specs.openstack.org/openstack/api-wg/guidelines/headers.html#avoid-proliferating-headers
13:25:24 <ruijie_> yes Qiming
13:26:22 <Qiming> also this: http://specs.openstack.org/openstack/api-wg/guidelines/http.html
13:26:38 <Qiming> we are supposed to return a proper status code for this
13:28:10 <Qiming> I don't think we should add an 'extra' field to each and every API request ...
13:28:15 <Qiming> it is gonna be dirty
13:30:04 <ruijie_> thanks Qiming, let's talk about it a little bit later :)
13:30:20 <Qiming> alright
13:30:50 <ruijie_> another one is about this patch https://review.openstack.org/#/c/504830/
13:31:06 <ruijie_> to revisit the documentation of event dispatcher
13:31:49 <ruijie_> the developer part of the event dispatcher looks good now？
13:32:46 <Qiming> yes I think so
13:33:33 <Qiming> however, I think we are supposed to split that rst file into two
13:33:37 <ruijie_> http://git.openstack.org/cgit/openstack/senlin/tree/doc/source/contributor/event_dispatcher.rst
13:33:45 <Qiming> one for user, one for developer
13:34:03 <ruijie_> this one is about how to extend the event plugin
13:34:36 <ruijie_> Qiming, you mean we may need to divid the user reference to 2 parts?
13:34:53 <Qiming> the contents under 'contributor' subdir are there for developers
13:35:11 <Qiming> the one in your patch is targeting users
13:35:40 <ruijie_> yes Qiming
13:35:51 <Qiming> I see, we already have guide for developers
13:36:10 <ruijie_> yes Qiming :)
13:36:51 <ruijie_> next one. the tempest-plugin
13:37:06 <ruijie_> https://review.openstack.org/#/c/504870/
13:37:39 <ruijie_> the patch is already there, will keep checking it
13:38:21 <ruijie_> that's all from my side :)
13:38:48 <Qiming> cool
13:39:00 <ruijie_> I noticed longkb_ is here :)
13:39:23 <Qiming> we will need to revise some gate jobs later when the github project is migrated
13:39:59 <ruijie_> yes Qiming, after that, will move the gate jobs from senlin to the plugin step by step
13:40:34 <Qiming> to be honest, I'm not sure
13:41:11 <Qiming> the gate jobs are there to ensure changes to senlin repo won't break anything, they are not there to protect the tempest plugin
13:41:58 <Qiming> so maybe the jobs are supposed to be revised to invoke tempest plugin project, and they still vote for senlin patches
13:43:00 <Qiming> also speaking of your event doc improvement
13:43:09 <Qiming> when reading this: https://docs.openstack.org/senlin/latest/user/events.html
13:43:43 <Qiming> I felt that we may need a subsection for the dispatcher configuration so that users will easily find the info
13:44:32 <ruijie_> one for the database and another for the messaging
13:44:54 <Qiming> we need a subsection in that article
13:45:06 <Qiming> for readability's purpose
13:45:19 <Qiming> just a quick thought
13:45:35 <ruijie_> okay, that makes sense
13:46:24 <ruijie_> next one is about the meetup
13:47:30 <ruijie_> we can organize the meetup in Beijing in the 2nd/3rd week of Oct
13:47:40 <Qiming> any suggestion on the date/location ?
13:48:30 <ruijie_> how about 15th Oct
13:49:30 <ruijie_> Queens-1 milestone :)
13:49:32 <Qiming> it is okay for me
13:49:54 <Qiming> I may have a conflicting schedule but not yet fixed
13:50:48 <ruijie_> the date could be discussed, will send an email to mailing list. I think the location could be Beijing :)
13:52:30 <ruijie_> :< >: that's all from me
13:54:11 <Qiming> no topic from me
13:54:36 <ruijie_> okay, we can end it a little bit early
13:54:46 <ruijie_> thanks for joining :)
13:55:04 <ruijie_> #endmeeting