14:19:56 <Li_Liu> #startmeeting openstack-cyborg 14:19:57 <openstack> Meeting started Wed Nov 28 14:19:56 2018 UTC and is due to finish in 60 minutes. The chair is Li_Liu. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:19:58 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:20:00 <openstack> The meeting name has been set to 'openstack_cyborg' 14:20:02 <Sundar> Do we have the meeting time on the agenda? 14:20:14 <Li_Liu> #topic Roll Call 14:20:24 <Li_Liu> #info Li_Liu 14:20:26 <Sundar> #info Sundar 14:20:31 <Coco_gao> When will we change the meeting time? 14:20:46 <Li_Liu> we will use a new meeting time next week 14:20:49 <Coco_gao> #info Coco 14:20:54 <Coco_gao> OK, got that. 14:21:07 <Sundar> We first need to agree on the new time. :) 14:21:20 <Li_Liu> #topic status update on Nova interaction spec 14:21:44 <Li_Liu> Sundar, for sure, I am still waitiing for more inputs 14:22:02 <shaohe_feng_> #info shaohe_feng_ 14:22:09 <Li_Liu> prob will wait till the end of this week 14:22:30 <Li_Liu> Sundar, could you provide some update on the nova spec so far? 14:24:07 <Sundar> The status is as it says in the spec review and in the IRC chats I pointed to. At a high level, a group of us at the PTG had hashed out a workflow but apparently not all Nova developers are on board with that. 14:24:44 <Sundar> Secondly, there are discussions on the API design and use of groups within device profiles. 14:25:00 <Coco_gao> OK, thank you for the update. 14:25:06 <Sundar> Further, on close examination of what os-acc does, 14:25:39 <Coco_gao> Did they reach an agreement that they don't need os-acc? 14:25:55 <Sundar> a group of us in the IRC said we don't need os-acc. We need some way to ensure that we are carrying a quorum of Noava developers with such deicsions 14:26:43 <Sundar> Part of what makes these decisions complicated is that we get requests to look at 'the code', but we don;t have working code that we can show to others 14:27:24 <Coco_gao> I think right now, everything is blocked on the new DB implementation 14:27:39 <Sundar> So, I am putting together a simple prototype. That is going to evolve a lot based on feedback. But some of the constituents, such as the db changes, can be leveraged hopefully 14:28:20 <Sundar> Coco: I have shared my db layer prototype with you. What do you think? 14:29:15 <Li_Liu> I am back.. irc jsut crashed 14:29:24 <Coco_gao> I will send you an email tomorrow to further discuss on that. 14:29:54 <Li_Liu> Sunder, do you need any help on that? 14:30:22 <Sundar> Li_Liu: help in ensuring that all Nova folks are aligned? What do you propose to do? 14:30:37 <Coco_gao> IMHO,everything seems OK, but I wonder whether we need that much of tables. 14:30:43 <Sundar> If it is help on the POC, 14:31:15 <Sundar> I am concerned it can be confusing because things change on a dime. However, I would appreciate feedback on the POC 14:31:55 <Li_Liu> can we leverage the POC from shaohe_feng a while ago/? 14:32:03 <Sundar> Coco: Got it. This is a fully normalized db schema. We can merge some tables together, but I don't know if that is wise 14:32:39 <Sundar> Li_Liu: This is a totally new flow -- we are not focused on the drivers/discovery etc. but on Nova calls, Cyborg APIs for ARQs etc. 14:32:45 <Coco_gao> Besides, there are difference in your implementation and Li's spec, I hope we can reach an agreement about the details and start coding. 14:32:47 <Li_Liu> Sundar, I believe some of them can totally be merged 14:33:15 <Sundar> Li_Liu: which tables? 14:33:45 <Li_Liu> Coco_gao, don't worry, Sundar and I will align the design and PoC 14:34:25 <Li_Liu> like AttachHandleMdev and AttachHandle 14:34:40 <Coco_gao> OK, I can start the db implementation this week 14:34:52 <Sundar> Can we state which tables are being discussed? The scheme needs to reflect our device model 14:34:52 <Li_Liu> Thanks a lot Coco_gao 14:35:21 <Coco_gao> Maybe we can first start with device table. 14:35:29 <Coco_gao> Do you think we need another meeting? 14:35:38 <Sundar> Li_Liu: That is like merging a parent class and a child class. If we do that, there will be redundant fields and logic in each child class. 14:36:28 <Li_Liu> I don't feel good to have sperated table for subclasses 14:36:50 <Coco_gao> Yes, I agree with Li 14:36:56 <Sundar> Coco: If you are ok, I can share the db code with Li and others 14:37:01 <Li_Liu> I have a feeling we might need a zoom meeting to finalized all of these.. 14:37:08 <Sundar> Sure 14:37:23 <Coco_gao> Yes, pls share the code. 14:37:27 <Li_Liu> For this IRC, let's not dive into tech details 14:38:11 <Li_Liu> Sundar, please share the code and let's find a time for zoom meeting to discuss the spec in details 14:38:14 <Coco_gao> OK. Zhenghao is not available these two week. 14:38:20 <Coco_gao> weeks 14:38:21 <Li_Liu> i c 14:39:26 <Li_Liu> #todo 1. Sundar shares the code with rest of the team 14:39:54 <Li_Liu> #todo 2. shcedule a zoom meeting to finalize the db spec 14:40:20 <Li_Liu> #todo 3. Sundar keeps working on the Nova spec 14:40:56 <Li_Liu> #todo 4. Coco can start the db implementation some time this week (better after the zoom meeting) 14:41:43 <Li_Liu> did I miss anything? 14:42:18 <Sundar> Who will schedule the zoom meeting? 14:42:25 <Li_Liu> I will 14:43:24 <Coco_gao> OK, my new driver patch is ready, but only the structure, not some functions. The function is depend on the new db schema . So it will be uploaded after the db patch, I think. 14:43:44 <Sundar> We can continue to discuss over email before the zoom meeting too 14:43:45 <Li_Liu> sure, thanks a lot Coco 14:43:51 <Li_Liu> yup 14:45:45 <Li_Liu> LI LIU is inviting you to a scheduled Zoom meeting. Topic: Cyborg DB scheme discussion Time: Nov 29, 2018 10:00 AM Eastern Time (US and Canada) Join Zoom Meeting https://zoom.us/j/488464582 One tap mobile +16465588665,,488464582# US (New York) +17207072699,,488464582# US Dial by your location +1 646 558 8665 US (New York) +1 720 707 2699 US Meeting ID: 488 464 582 Find your local number: https://zoom.us/u 14:46:09 <Li_Liu> tomorrow around this time? is that ok? 14:46:33 <Li_Liu> 11 pm beijing time/7 am pst 14:46:41 <Sundar> I'd prefer a time that is morning China and evening pacific 14:47:00 <Li_Liu> ah 14:47:06 <Coco_gao> I prefer Sundar's proposal. 14:47:09 <Li_Liu> does morning work for you coco? 14:47:11 <Sundar> Better for everybody? 14:47:31 <Li_Liu> sure thing, will move to a later time 14:47:33 <Coco_gao> Morning is always OK for me. 14:48:41 <Li_Liu> #endmeeting