03:02:09 <Li_Liu> #startmeeting openstack-cyborg
03:02:10 <openstack> Meeting started Wed Jan  9 03:02:09 2019 UTC and is due to finish in 60 minutes.  The chair is Li_Liu. Information about MeetBot at http://wiki.debian.org/MeetBot.
03:02:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
03:02:13 <openstack> The meeting name has been set to 'openstack_cyborg'
03:02:20 <Li_Liu> #topic Roll Call
03:02:23 <wangzhh> Hi all.
03:02:25 <Li_Liu> #info Li_Liu
03:02:34 <Li_Liu> hi wangzhh
03:02:39 <wangzhh> #info wangzhh
03:02:40 <Yumeng_> #info Yumeng_
03:02:48 <Yumeng_> hi
03:03:02 <Li_Liu> Let's wait for others for a bit
03:03:05 <xinranwang> #info xinranwang
03:03:08 <xinranwang> hi all!
03:03:24 <Yumeng_> hi xinranwang
03:03:57 <Li_Liu> Hi Xinran
03:04:11 <Li_Liu> wangzhh, do you know if coco is coming?
03:04:33 <wangzhh> Let me ping her.
03:05:47 <jiapei> #info jiapei
03:06:00 <Coco_gao> Hi all
03:06:11 <wangzhh> Welcome.
03:06:55 <Li_Liu> ok, let's get started
03:07:20 <Li_Liu> #topic coco's patches on DB scheme
03:07:29 <Coco_gao> #info Coco_gao
03:07:36 <Li_Liu> #link  https://review.openstack.org/#/c/625630/
03:07:41 <Coco_gao> I will update tonight.
03:08:10 <Coco_gao> Add some unit test
03:08:20 <Li_Liu> maybe ask sundar to give some comments?
03:08:46 <Li_Liu> but I know sundar is on vacation travelling.
03:09:20 <Li_Liu> I will ping him on wechat to see if he has time
03:09:27 <Coco_gao> Yeah, also,  we should first update the unit test.
03:09:57 <Coco_gao> If not, the code can't pass.
03:10:00 <Li_Liu> do you wanna do the unit test in this patch or a new patch?
03:10:24 <Coco_gao> At least, I should delete the old unit test in this patch.
03:10:59 <Li_Liu> ok, I am fine with ether way
03:11:24 <Coco_gao> ok. I will see if time is enough for the new unit test.
03:11:47 <Coco_gao> What about other people's object ?
03:11:53 <Coco_gao> ovo object
03:12:07 <Li_Liu> I have the confirmed with release team. the ms2 is on Jan 10
03:12:24 <Li_Liu> but it's ok to push it back a bit
03:12:52 <Coco_gao> Hi, shaohe
03:12:53 <shaohe_feng_> #info shaohe
03:13:03 <shaohe_feng_> Morning Coco_gao
03:13:10 <Yumeng_> coco, I will update the db change in cyborg/db/sqlalchemy/api.py, so you can ignore any changes in cyborg/db/sqlalchemy/api.py. I will do that
03:13:14 <wangzhh> morning shaohe
03:13:28 <shaohe_feng_> wangzhh: morning.
03:13:43 <xinranwang> hi shaohe_feng_
03:14:19 <shaohe_feng_> xinranwang: hi.
03:14:31 <shaohe_feng_> what's the process of DB refactor
03:14:33 <Coco_gao> thanks yumeng. Does api change contains the object part?
03:14:55 <Li_Liu> hi shaohe
03:14:57 <wangzhh> https://review.openstack.org/#/c/596691/  Could u help to merge this patch.
03:15:31 <Li_Liu> Coco_gao: the ovo code is some what depends on the db patch i believe
03:15:43 <shaohe_feng_> Li_Liu: how is your vocation?
03:15:58 <Coco_gao> yes, I will merge the db part first.
03:16:10 <Li_Liu> shaohe_feng_: busy with the kid.... no fun :(
03:16:11 <Yumeng_> I suppose yes, at least accelerator object
03:16:35 <Coco_gao> but everyone can start their jobs first according to Sundar's patch.
03:16:38 <Yumeng_> coco:I suppose yes, at least accelerator object
03:17:08 <Coco_gao> After my patch, they are merge their own.
03:17:17 <Coco_gao> they can
03:17:36 <Li_Liu> Coco_gao: I remember you will be working on deployable and attribute after this one right?
03:18:06 <Coco_gao> Because it won't be done in a short time, so we'd better start working first, than think about the merge thing.
03:18:14 <Coco_gao> Li_liu, yes
03:18:22 <Coco_gao> that's my job.
03:18:32 <Li_Liu> wangzhh, do you wanna help with planing the rest of ovo work? you don't need to do them all.
03:18:53 <Li_Liu> you can appoint some pieces to me and xinran
03:20:17 <wangzhh> OK,  let me assign them on storyboard.
03:20:32 <Li_Liu> great
03:20:38 <Coco_gao> Hi, changzhi
03:20:46 <changzhi> hi
03:21:00 <Li_Liu> please send a email to the rest of the team once you done it wangzhh
03:21:01 <xinranwang> yes, please let me know when you need help. :)
03:21:23 <Li_Liu> ok, let's move on
03:21:39 <wangzhh> Got it.
03:21:54 <Li_Liu> since sundar is not here today, I will have a separated discussion with him afterwards
03:22:06 <Li_Liu> about his POC work
03:22:34 <Li_Liu> do you guys have any suggestion/comments on the PoC patches from Sundar, please help review them
03:24:13 <Li_Liu> #topic AoB
03:25:12 <Li_Liu> Hi Sundar
03:25:23 <Sundar> Hi Li
03:25:31 <Li_Liu> Happy New year
03:25:53 <Li_Liu> I have just wrapped up with other topics
03:25:57 <Sundar> Happy New Year to you too!
03:25:59 <Sundar> Sorry, I am late. My flight arrived just now.
03:26:08 <Li_Liu> do you wanna quickly give a update on your side?
03:26:12 <Li_Liu> it's ok
03:26:30 <Sundar> The Cyborg and CYborg client code are both in their feature branches.
03:26:57 <shaohe_feng_> Sundar: Happy New Year.  You are on a travel.
03:27:01 <Sundar> Instead of using drivers/agent etc. to discover devices, I just populate the db directly.
03:27:30 <Sundar> Similarly, the Placement is populated directly too. That is done in the initial_setup script.
03:27:49 <Sundar> Shaohe_feng: Happy New Year.
03:28:29 <Li_Liu> Sundar, how are you gonna submit the feature patch for nova?
03:28:34 <Sundar> You can use curl or the cyclient script to access the APIs, to get device profiles, create/delete ARQs and bind/ubind them.
03:29:23 <Sundar> I have Nova patches that use Cyborg client to use these APIs. SO, it gets the device profile groups etc.
03:29:50 <Li_Liu> Is the nova patches in your feature branch already?
03:29:55 <Sundar> The Nova scheduler does not support nRPS well yet -- so the results of Placement query will not include the FPGA.
03:30:20 <Sundar> I work around that by just getting the selected compute node and grabbing the nested RP within that.
03:31:09 <Sundar> The libvirt driver patch then gets the ARQs, grab the attach handle within, and compose the VM xml. We can see the PCI device in the XML.
03:31:46 <Sundar> That's the broad outline. I'll stop for any questions now.
03:32:03 <Li_Liu> just to get a clarification tho, the feature branch's DB design does not need to align with the DB spec you merged late last year right?
03:32:24 <Sundar> Li_Liu: The Nova patches will be submitted directly to Nova master.
03:33:03 <Sundar> Li_Liu: yes. I replied to your comment. The pilot does not aim to be complete -- if I write everything, Coco and others don;t have much left :)
03:33:32 <Li_Liu> Sundar, i see
03:33:44 <Li_Liu> I will keep review them then
03:34:29 <Sundar> Yes, please do review the rest. Outside the db code, if there are any gaps not marked with TODO or FIXME, please LMK.
03:35:14 <Li_Liu> sure thing
03:35:53 <Li_Liu> folks, please help review Sundar's patches if you have the chance.
03:36:58 <Li_Liu> Thanks a lot for the work Sundar and Coco_gao :)
03:37:32 <Li_Liu> Any other things you guys wanna talk about?
03:37:38 <Sundar> Welcome :)
03:38:22 <Li_Liu> ok, let's call it
03:38:43 <Li_Liu> have a good night/day where ever you are in the world.
03:38:47 <Li_Liu> #endmeeting