03:13:56 #startmeeting openstack-cyborg 03:13:57 Meeting started Wed Dec 19 03:13:56 2018 UTC and is due to finish in 60 minutes. The chair is Li_Liu. Information about MeetBot at http://wiki.debian.org/MeetBot. 03:13:58 #info Coco_gao 03:13:59 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 03:14:01 The meeting name has been set to 'openstack_cyborg' 03:14:02 just on time CoCo 03:14:11 #topic Roll Call 03:14:16 #info wangzhh 03:14:16 #info Li_Liu 03:14:38 #info xinranwang 03:15:22 #info Sundar 03:15:24 Hi all 03:15:30 yummy is a story.. One of my classmate from Kuwait found it's really diffcult to pronounce Yumeng, but easy to pronunce Yummy.... then my name goes to Yummy... 03:15:36 #topic DB spec 03:15:42 Hi sundar. 03:15:49 https://review.openstack.org/#/c/615462/ 03:16:03 I addressed coco comments last night 03:16:25 If there are no more comments for another 24 hours, I will let it merge 03:17:21 OK. 03:18:16 Sundar, does it get unblocked/abandoned? 03:18:16 https://review.openstack.org/#/c/602978/ 03:19:00 Li_Liu: I am fine with it. We already have a disclaimer that small changes are ok 03:19:25 you mean this one ? https://review.openstack.org/#/c/615462/ 03:19:42 Device profiles are very much in place. The feature branch code is implementing it 03:19:59 Li_Liu: Yes, I was referring to db spec. 03:20:24 oh, how about this one? https://review.openstack.org/#/c/602978/ 03:20:29 it's from you 03:20:38 about device profiles. 03:21:27 Yes, as I sadi above, device profiles are part of our proposal, and feature branch code is implementing it. It may be revised slightly based on Nova feedback but the concept of device profiles will be there 03:21:32 *said 03:22:25 I see, so it will keep hanging there for now then 03:22:52 But I think we can start implementation based on Li's Spec right now, maybe not wait for nova feedback 03:23:03 Li_Liu: yes 03:23:12 Coco_gao: Agreed 03:23:23 that's right 03:23:33 #topic DB implementation 03:23:41 CoCo and Sundar 03:23:42 https://review.openstack.org/#/c/625630/ 03:23:51 https://review.openstack.org/#/c/625285/ 03:24:05 I see both patch has migration code 03:24:25 Li_Liu: They are in different branches 03:24:39 ah 03:24:45 My code is based on Sundar's work. 03:25:00 Coco_gao, wangzhh and Yumeng will do it for master branch 03:25:13 cool 03:25:32 hey are doing the 'proper' way. I may have many TODOs because I am trying to do just enough to get a working system that can be shown to Nova 03:25:51 Thanks a lot Sundar 03:25:59 Yes, we are foucing on different areas. 03:26:36 Li_Liu: Welcome 03:26:56 I will also submit objects layer code, that Coco and others can leverage 03:27:08 Thanks a lot 03:27:25 I also have working API layer code. I would suggest not to put that in master branch till we align with Nova 03:27:27 that will be much faster to get the work done 03:27:42 Coco_gao: Welcome. 03:28:02 Sundar, do you have anyone helping you on that? 03:28:08 seems quite some work 03:28:22 Also, in our engineering sync on Tuesday China time, we said that we should v1 APIs. But should we keep v1 objects and db layer? 03:28:49 Li_Liu: No :) 03:29:09 I started on Thanksgiving weekend and hope to finish on Christmas holiday 03:29:57 ah, speaking of that, we will not have IRC meetings for the next 2 weeks 03:30:10 due to holiday seasons 03:30:13 Yes, I think so. 03:30:33 Coco_gao and others: Do you have Christmas holidays in China? 03:30:46 Merry Christmas and Happy new year~ 03:31:08 I don't think so 03:31:18 Coco_gao and all: Same to you all! 03:31:19 there are holidays for X'mas 03:31:20 No, maybe just an hours holiday. :( 03:31:46 wangzhh: An hour's holiday? 03:31:50 No, we don't have holiday but usually we celebrate that~ 03:32:38 Cool. We said that we should v1 APIs. But should we keep v1 objects and db layer? 03:33:00 Sundar, Leave work early that day. 03:33:30 If v1 API is used to query an accelerator object, for example, we can return null list, because there are no accelerator tables 03:33:38 do the versions have to line up? 03:33:44 But, if v1 API is used to POST an accelerator, what's the right behavior? 03:34:00 invalid operation? 03:34:54 Li_Liu: I am not sure how OpenStack core defines backwards compat. 03:35:25 If we returned a valid object in Rocky, can we return invalid op in Stein for the same version? 03:37:23 I can ask around for that one 03:37:27 no worries 03:37:31 Thanks 03:37:59 I am not a fan of keeping outdated code around. It would be good to remove v1 objects 03:38:12 me too, ahhha 03:38:21 yup 03:38:34 that will be easier for implementation. 03:39:04 Yes. I just submitted a patch to the feature branch to disable all unit tests for v1 :) 03:39:21 #topic work assignments 03:39:48 Sundar, you will keep working on that pilot branch. I guess that will keep you busy for some time 03:40:08 Li_Liu: Absolutely 03:40:37 CoCo, Zhenghao, Yumeng, you guys will be focusing on the db implementations. and I will also be helping 03:42:00 Yup 03:42:15 That's it then 03:42:21 #topic AoB 03:42:41 any another thing you guys wanna share before the end of 2018? :P 03:42:59 Pls help review the code. 03:43:03 thank you 03:43:30 Please review the pilot branch code too 03:43:41 Early feedback is appreciated 03:43:44 That's for sure 03:43:50 I will, thanks Sundar 03:43:56 will do that :) 03:44:20 Thank you guys Mary X'Mas and happy new year!! 03:44:24 Thanks! Happy Holidays to all of you! 03:44:28 Thanks a lot for the work in 2018 03:44:35 let's keep pushing in 2019 03:44:56 Yup 03:45:12 Three cheers to that 03:45:27 Have a good night/day where ever you are 03:45:34 #endmeeting