13:00:02 <yanyanhu> #startmeeting senlin
13:00:03 <openstack> Meeting started Tue Nov  1 13:00:02 2016 UTC and is due to finish in 60 minutes.  The chair is yanyanhu. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:06 <openstack> The meeting name has been set to 'senlin'
13:00:16 <yanyanhu> hi
13:00:18 <elynn> hi
13:00:23 <lvdongbing> hi
13:00:24 <lixinhui> hi
13:00:31 <yanyanhu> hi, guys
13:00:54 <yanyanhu> lets wait for while for other attenders
13:01:50 <yanyanhu> ok, lets start, I think Qiming will join soon
13:01:59 <yanyanhu> here is the agenda https://wiki.openstack.org/wiki/Meetings/SenlinAgenda#Agenda_.282016-11-01_1300_UTC.29
13:02:10 <yanyanhu> please feel free to add any item you want to discuss
13:02:20 <yanyanhu> #topic ocata work items
13:02:29 <Qiming> o/
13:02:30 <yanyanhu> https://etherpad.openstack.org/p/senlin-ocata-workitems
13:02:32 <yanyanhu> hi, Qiming
13:02:51 <yanyanhu> ok, the first item
13:02:59 <yanyanhu> performance test. no progress recently
13:03:14 <yanyanhu> will work on it later after versioned request job is done
13:03:29 <yanyanhu> next one is HA
13:03:45 <yanyanhu> hi, Qiming, lixinhui
13:03:56 <yanyanhu> I guess no code work in last two weeks
13:04:02 <lixinhui> yes
13:04:10 <lixinhui> we delivered Auto-healing
13:04:10 <yanyanhu> but do have some new ideas/plan after summit :)
13:04:15 <lixinhui> presentation
13:04:18 <yanyanhu> cool
13:04:33 <lixinhui> and get good feedback there, from masakari
13:04:34 <yanyanhu> how is it going
13:04:47 <yanyanhu> taker team?
13:04:55 <lixinhui> the HA team
13:04:59 <lixinhui> and Adam
13:05:01 <yanyanhu> ah, I see
13:05:13 <lixinhui> the suse guy
13:05:27 <lixinhui> and one pacemaker guy
13:05:50 <lixinhui> they suggest we cooperate on HA spec and integration together
13:05:58 <yanyanhu> great
13:06:15 <lixinhui> on my side, I may draft a BP about mistral integration with Senlin
13:06:21 <lixinhui> in this week
13:06:33 <lixinhui> will need your comments and suggestions
13:06:44 <yanyanhu> nice, please ping us in the channel and will definitely help to review
13:06:59 <lixinhui> :)
13:07:03 <yanyanhu> I believe there are lots of jobs we can do together with them
13:07:07 <elynn> And they also suggest us to attend their weekly meeting.
13:07:19 <lixinhui> Yes :)
13:07:31 <lixinhui> Thursday afternoon?
13:07:33 <yanyanhu> yes, I saw the message you left in the etherpad
13:07:46 <yanyanhu> our afternoon?
13:08:04 <yanyanhu> hope that won't conflict with your day time work
13:08:18 <elynn> Can't remember the time
13:08:23 <yanyanhu> I think Qiming joined that meeting before
13:08:43 <elynn> https://wiki.openstack.org/wiki/Meetings/HATeamMeeting
13:08:55 <yanyanhu> ok, we can check the schedule later. Hope you and Qiming can get time to join it, lixinhui
13:08:57 <elynn> Can't find meeting time info from above page...
13:09:04 <lixinhui> Okay
13:09:12 <lixinhui> will find the time later
13:09:17 <yanyanhu> elynn, I recall its our afternoon but not very sure
13:09:25 <yanyanhu> lixinhui, thanks :)
13:09:29 <lixinhui> :)
13:09:56 <yanyanhu> ok, plz update the etherpad if there is any new progress about HA, lixinhui Qiming, :)
13:09:58 <yanyanhu> thanks
13:10:08 <lixinhui> sure
13:10:10 <yanyanhu> ok, lets move on
13:10:16 <yanyanhu> document
13:10:17 <elynn> Ah, here http://eavesdrop.openstack.org/#High_Availability_Meeting Monday afternoon.
13:10:26 <yanyanhu> elynn, nice
13:10:35 <yanyanhu> maybe I will join it as well
13:10:39 <yanyanhu> as listener :)
13:11:08 <yanyanhu> ok, next one
13:11:17 <yanyanhu> versioned requests
13:11:26 <yanyanhu> Qiming is now focusing on it
13:11:34 <yanyanhu> and I'm also trying to provide some help
13:11:44 <yanyanhu> the basic framework has been ready before summit
13:12:11 <yanyanhu> so the current job is migrating old non-versioned implementation to versioned one
13:12:31 <yanyanhu> Qiming is working on cluster related APIs and I'm now working on node related ones
13:12:56 <yanyanhu> you guys can refer to our recent patches to get more idea about this work
13:13:08 <yanyanhu> and welcome more hands on it if you guys have time :)
13:13:26 <yanyanhu> this the one of the most important work we want to finish this cycle
13:13:49 <yanyanhu> and without it, it will be risky to make any change on our API interfaces
13:13:56 <yanyanhu> and also rpc interface I think
13:14:04 <yanyanhu> so hope can finish it soon
13:14:26 <yanyanhu> hi, Qiming, any more comment on the detail?
13:15:08 <yanyanhu> ok, lets move on
13:15:14 <yanyanhu> zaqar message receiver
13:15:28 <yanyanhu> spent some time in last two weeks to address some backlogs
13:15:41 <yanyanhu> including add more support to openstacksdk
13:15:54 <yanyanhu> brian have helped to review and approve those patches
13:16:06 <yanyanhu> and I think message receiver can now work correctly
13:16:20 <yanyanhu> I'm still working on integration test for it and this job has almost been done
13:16:35 <yanyanhu> only need to wait for sdk to bump its version to make it work
13:16:54 <yanyanhu> after that, I will consider to build a story as reference for user
13:17:15 <yanyanhu> to illustrate how to use this new type of receiver
13:17:29 <briancurtin> yanyanhu: i’m actually doing a release right now so a new version should be out within an hour. would have been out right now but got busy at the end of yesterday
13:18:01 <yanyanhu> briancurtin, great, thanks a lot :)
13:18:27 <yanyanhu> I will try to check the integration test after new release is ready
13:19:03 <yanyanhu> ok, any comments or question about this work?
13:19:27 <yanyanhu> ok, next one
13:19:37 <yanyanhu> event/notification, no progress I guess
13:19:55 <yanyanhu> next one, batch policy
13:20:05 <yanyanhu> ruijie is not here?
13:20:25 <yanyanhu> ruijie is now working on it
13:20:42 <yanyanhu> the basic support has been done and he is now adding batch policy support to more cluster actions
13:20:47 <yanyanhu> like cluster_delete
13:20:52 <yanyanhu> node_delete, etc.
13:21:22 <yanyanhu> please help to review the code if you guys have time :)
13:21:24 <yanyanhu> thanks
13:21:38 <elynn> okay, will
13:21:45 <yanyanhu> ok, above is all items in the etherpad
13:22:32 <yanyanhu> hi, XueFengLiu lvdongbing, plz feel free to add work items you are working on to the etherpad as well :)
13:22:50 <lvdongbing> Ok :)
13:22:54 <yanyanhu> that will be helpful for us to track them
13:23:02 <yanyanhu> thanks :)
13:23:13 <yanyanhu> I believe you guys are also working on some important items
13:23:39 <yanyanhu> although they are not on the list
13:24:03 <yanyanhu> ok, next item in agenda
13:24:26 <yanyanhu> #topic news from Barcelona summit
13:24:27 <Qiming> xuefeng is trying to rework the senlin lock thing
13:24:35 <yanyanhu> Qiming, ah right
13:24:39 <Qiming> but I am not sure he is changing it in the right way
13:24:43 <yanyanhu> the first patch has been there
13:25:01 <yanyanhu> Qiming, maybe this worth a spec/bp?
13:25:11 <Qiming> don't think so
13:25:18 <yanyanhu> I think we'd better make a more thorough discussion on it before starting code job
13:25:26 <Qiming> need some explanation why this change is needed
13:25:29 <yanyanhu> yes
13:25:42 <Qiming> yes, or else it would be a waste of time
13:26:16 <yanyanhu> xuefeng is not online now
13:26:31 <yanyanhu> maybe we can have a discussion on this topic in irc tomorrow
13:26:37 <Qiming> can sync with him later
13:26:42 <yanyanhu> ok
13:27:06 <yanyanhu> so if no update for work item, maybe you guys can give us some update about summit :)
13:27:13 <yanyanhu> news
13:27:18 <yanyanhu> Qiming, lixinhui elynn :)
13:27:42 <Qiming> okay
13:27:54 <Qiming> we delivered the two presentations, well done
13:28:16 <yanyanhu> nice
13:28:18 <Qiming> particularly the VM HA story, well received by the community wide vm-ha story team
13:28:38 <Qiming> senlin may play an important role in that story
13:28:50 <Qiming> and it is now an expectation from product working group
13:28:54 <yanyanhu> cool.
13:29:19 <Qiming> Adam et al are working on a protocol for VM HA events so that different OpenStack components can be easily integrated
13:29:28 <yanyanhu> do need more cross-project effort to build the HA solution
13:29:35 <Qiming> I asked him to leave a message when there is a draft for review
13:29:58 <yanyanhu> yes, that will very important for our listener
13:30:06 <Qiming> they have openstack-ha meetings every Thursday afternoon (need double check), they want us to join it
13:30:14 <yanyanhu> if there is a standard for the format, that will be very helpful
13:30:46 <Qiming> could be Monday ...
13:30:55 <yanyanhu> http://eavesdrop.openstack.org/#High_Availability_Meeting
13:31:00 <yanyanhu> yes, maybe Monday
13:31:06 <yanyanhu> elynn just found this link
13:31:23 <yanyanhu> will join it as well as listener
13:31:35 <Qiming> time is not bad for UTC+8
13:31:41 <yanyanhu> yup
13:31:56 <Qiming> next thread is about integration with Magnum
13:32:18 <briancurtin> yanyanhu: sorry to interrupt, but just wanted to say openstacksdk 0.9.9 is now available: https://pypi.python.org/pypi/openstacksdk/0.9.9
13:32:24 <Qiming> I joined their design session, brought up this as an alternative to their new clustering feature
13:32:34 <yanyanhu> briancurtin, great!
13:32:48 <yanyanhu> thank you so much for this quick verion bumping
13:33:00 <Qiming> let's see if gate is using the latest version of sdk
13:33:01 <yanyanhu> will test it tomorrow after I go back to office
13:33:10 <yanyanhu> Qiming, yes
13:33:44 <Qiming> magnum team has some special requirements that they cannot get from Heat, seems falling completely into senlin's mission
13:34:20 <Qiming> Ton from the magnum team will have a call with me to understand the detailed requirements next week
13:34:32 <yanyanhu> Qiming, yes, I noticed their discussion about node group in mailing list
13:34:39 <Qiming> the third item is about integration with Tacker
13:34:43 <yanyanhu> nice, will jion it as well
13:34:49 <yanyanhu> s/jion/join
13:34:55 <Qiming> Haiwei proposed a session with tacker time to start this
13:35:17 <elynn> Could you add me in that call? :)
13:35:18 <Qiming> seems auto-scaling and auto-heaing are both features they want
13:35:30 <yanyanhu> elynn, will let you know if the time is decided
13:35:47 <Qiming> sure, elynn, don't have the invitation yet, will forward
13:36:03 <elynn> Thanks Qiming yanyanhu
13:36:14 <lixinhui> me2 plus
13:36:21 <Qiming> they don't have a complete understanding about senlin, so we gave them an intro
13:36:25 <yanyanhu> sure
13:36:27 <Qiming> okay, lixinhui
13:36:38 <lixinhui> :)
13:36:43 <Qiming> the primary concern is about the invocation path:
13:36:58 <Qiming> TOSCA->heat-translator->heat-template
13:37:19 <Qiming> that means two things: they have to use Heat for consuming TOSCA input
13:37:44 <Qiming> and they have to use senlin indirectly through heat templates
13:38:12 <Qiming> however, after confirmation with Sadhev, neither one should be a barrier
13:38:29 <Qiming> heat-translator can be easily adapted to generate senlin profile/policy specs directly
13:38:41 <yanyanhu> good news
13:38:47 <Qiming> although the implementation is not there yet
13:39:09 <yanyanhu> I believe there should be no barrier here for we can have lots of discussion with TOSCA team last year
13:39:24 <yanyanhu> we already have
13:39:35 <Qiming> then they understood ... for single stack cosisting of a combination of many different type of resources, they will continue to use Heat
13:39:57 <Qiming> when they need to build a pool of resources, they will turn to senlin
13:40:35 <Qiming> there may be some additional requirements for scaling pools of resources at a finer granularity
13:40:48 <yanyanhu> yes, they can keep using heat as the orchestration engine. if they have requirement for resource pool, senlin cluster is a good choice
13:40:54 <Qiming> doesn't look like a barrier though
13:41:02 <yanyanhu> yes, won't be
13:41:14 <Qiming> yes, we have more knobs for them to manage such pools
13:41:22 <yanyanhu> right
13:41:32 <yanyanhu> that why we started this project
13:41:35 <Qiming> that was the 3rd item
13:41:46 <Qiming> the next item is about baremetal support
13:42:21 <Qiming> some Huawei people joined out morning session, raising requirements on manageing edge devices, which are baremetal machines
13:43:04 <Qiming> we shared the existing BP (from yongli) and expressed our expectation to resource investments
13:43:25 <yanyanhu> https://blueprints.launchpad.net/senlin/+spec/scaling-bearmetal
13:43:26 <Qiming> the gentleman promised that they will invest some people to get this done
13:43:32 <elynn> I think we can find more from this link for what they are doing. https://etherpad.openstack.org/p/meghdwar
13:43:34 <yanyanhu> hope somebody can pick it up :)
13:44:15 <Qiming> I have just marked the BP as obsolete because we are not receiving requirements on it
13:44:27 <Qiming> need to resume it now
13:44:34 <yanyanhu> elynn, thanks for the link :)
13:44:40 <yanyanhu> Qiming, yes
13:45:02 <Qiming> seems lixinhui and haiwei are both interested in the NFV story
13:45:11 <yanyanhu> actually baremetal cluster is a very common requirement for many use cases I think
13:45:17 <Qiming> maybe we can create a thread for that
13:45:22 <yanyanhu> sure
13:45:34 <yanyanhu> will add this item to etherpad as well
13:45:39 <yanyanhu> to make it easy to track
13:45:42 <Qiming> lixinhui, please amend if I'm missing anything
13:46:09 <lixinhui> no, your introduction is complete and perfect :)
13:46:18 <yanyanhu> have the same feeling, haha
13:46:37 <Qiming> such a nice girl, never say things bad about me
13:46:45 <lixinhui> :)
13:46:53 <elynn> I think nova can manage bare metal, why do we need a bare metal cluster?
13:46:59 <yanyanhu> LoL
13:47:22 <yanyanhu> elynn, maybe manging it through ironic is a better way
13:47:26 <Qiming> managing a server is not the same thing as managing a cluster, right?
13:48:07 <elynn> you mean we might need some special policy for bare metal ?
13:48:21 <Qiming> we need profile type
13:48:56 <elynn> a profile type to use ironic ?
13:49:03 <yanyanhu> yes, think so
13:49:39 <yanyanhu> and maybe for other barematel backends
13:49:46 <yanyanhu> if needed
13:50:36 <Qiming> their api is here: http://developer.openstack.org/api-ref/baremetal/
13:50:41 <elynn> I see ..
13:50:50 <Qiming> you cannot treat it just as a different nova
13:52:17 <yanyanhu> ok, looks good
13:52:25 <yanyanhu> any more comments here?
13:52:33 <elynn> I thought ironic is a driver for nova, seems I was wrong.
13:52:48 <yanyanhu> elynn, it is also a individual service I think
13:53:00 <yanyanhu> just nova has virt driver for it
13:53:04 <yanyanhu> I guess
13:53:22 <Qiming> I think xuefeng is online, but he doesn't know how to join a channel, ...
13:53:32 <yanyanhu> ok...
13:53:42 <yanyanhu> I saw he dropped
13:53:51 <yanyanhu> and didn't join back
13:54:11 <yanyanhu> anyway, we have 6 mins left
13:54:22 <Qiming> well, he left a message on code review , :D
13:54:33 <yanyanhu> nice
13:54:36 <yanyanhu> will check it
13:54:45 <yanyanhu> #topic open discussion
13:54:47 <Qiming> Patch Set 1:
13:54:47 <Qiming> I am online.But I can't access #openstack-meeting now.
13:54:53 <yanyanhu> ...
13:55:05 <yanyanhu> maybe network problem
13:55:33 <yanyanhu> the most interesting way to leave msg I've ever seen :)
13:56:05 <yanyanhu> ok, we have 4mins
13:56:10 <yanyanhu> any more topics?
13:56:18 <Qiming> no from me
13:56:28 <yanyanhu> hi, xuefengliu, good to see U again :)
13:57:02 <yanyanhu> ok, if not further item to discuss, we can release the channel a little bit earlier
13:57:05 <yanyanhu> this time
13:57:16 <Qiming> ok
13:57:19 <xuefengliu> use phone now
13:57:23 <yanyanhu> ...
13:57:28 <yanyanhu> network is unstable
13:57:45 <xuefengliu> network has some problem
13:57:45 <yanyanhu> ok, thanks all you guys for joining
13:57:52 <yanyanhu> xuefengliu, yes...
13:58:12 <yanyanhu> will end the meeting. Have a good night :)
13:58:22 <Ruijie_> have a good one
13:58:27 <yanyanhu> :)
13:58:33 <Qiming> bbl
13:58:38 <yanyanhu> 88
13:58:45 <elynn> 8
13:58:45 <yanyanhu> #endmeeting