03:05:59 <Yumeng> #startmeeting openstack-cyborg 03:06:01 <openstack> Meeting started Thu Apr 9 03:05:59 2020 UTC and is due to finish in 60 minutes. The chair is Yumeng. Information about MeetBot at http://wiki.debian.org/MeetBot. 03:06:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 03:06:02 <chenke> Hi all. sorry for later. 03:06:04 <openstack> The meeting name has been set to 'openstack_cyborg' 03:06:04 <chenke> #info chenke 03:06:14 <xinranwang> #info xinranwang 03:06:17 <Yumeng> #topic Roll call 03:06:21 <Yumeng> #info Yumeng 03:06:33 <s_shogo> #info s_shogo 03:07:28 <Yumeng> I suppose the agenda for today is reviewing pending patches 03:08:07 <Yumeng> let's go through by priorities from high 03:08:33 <xinranwang> The microverison patch has been merged, thank you all for reviewing :) 03:08:47 <chenke> good job. 03:09:56 <Yumeng> great! Thanks Xinran for the efforts ! and v2 doc need review https://review.opendev.org/#/c/715980/ 03:10:55 <xinranwang> Yes, please review this doc patch as well. thanks Yumeng 03:11:34 <Yumeng> s_shogo and xinranwang, do you have any other patches need review? 03:11:58 <xinranwang> No other patches from m 03:12:00 <xinranwang> *me 03:12:06 <Yumeng> ok 03:12:30 <openstackgerrit> YumengBao proposed openstack/cyborg master: Refactor v2 arq api https://review.opendev.org/696089 03:12:55 <s_shogo> This one. I'm working with Sundar's comment. https://review.opendev.org/#/c/698190/ 03:13:56 <chenke> Ok. will reivew it. 03:14:04 <s_shogo> Shaohe may have opinion about this programming patch, as the previous meeting : I can ask him in wechat. 03:14:12 <s_shogo> thank you, chenke 03:14:37 <Yumeng> I notice shaohe also has several patches realted to programming https://review.opendev.org/#/q/owner:shaohe.feng%2540intel.com+status:open 03:15:17 <Yumeng> are they related to yours? s_shogo 03:16:14 <s_shogo> At a glance, It seems to be driver layer, my patch provides api. 03:16:28 <s_shogo> I'll investigate and check them after this meeting, thanks. 03:16:54 <Yumeng> ok. seems different. I will ask him if need merged in this release. 03:17:21 <s_shogo> ok. 03:17:34 <Yumeng> so there also left Brin's patches https://review.opendev.org/#/q/owner:zhangbailin%2540inspur.com+status:open 03:18:49 <Yumeng> and mine two: arq refactor https://review.opendev.org/#/c/696089/, and add policy role https://review.opendev.org/#/c/699102/ 03:19:31 <Yumeng> please help to review! not big change. 03:20:15 <Yumeng> chenke: do you have any pending patches? 03:20:37 <chenke> oh. no patch for me. But a patch needs us wo notice. 03:20:37 <xinranwang> Ok, will review them. 03:20:44 <chenke> https://review.opendev.org/#/c/718584/1/cyborg/conductor/manager.py 03:21:31 <chenke> This patch seems help cyborg to sync data with placement. We should consider whether it's needed or not. 03:23:54 <xinranwang> Yes, I just saw this. 03:24:38 <xinranwang> IMO, it is reasonable. 03:25:45 <Yumeng> +1 agree with Brin's comment. xinranwang: do we plan to decouple placement report in next release? 03:26:34 <chenke> It looks okay, but needs further analysis. 03:27:17 <xinranwang> I think it depends on users' demand, if we want cyborg run as a standalone project, yes, we should decouple them 03:29:42 <chenke> It looks okay, but needs further analysis. To be honest, the current piece of code looks a bit redundant. About placement reporting and cyborg database update are mixed together. But I have no better idea about decoupling them. 03:30:50 <xinranwang> chenke: yes, that's what I am thinking too. If we want decouple them, it will generate more redundant code, because both of them need to do the diff. 03:32:46 <chenke> Ye. seems no better code to make report and sync cyborg db. Let's these code stay as it is. 03:32:50 <Yumeng> emmm,seems a problem to choose a less evil one. let's keep it as one topic of the PTG. 03:33:26 <Yumeng> #topic AoB 03:33:35 <Yumeng> Anything else? 03:33:35 <brinzhang> https://review.opendev.org/#/c/718584/1/cyborg/conductor/manager.py are you talking this patch? 03:33:54 <Yumeng> yes brinzhang 03:34:24 <xinranwang> Yumeng: Sure, btw, do we have ptg etherpad where we can put the topic in, I'd like add some :) 03:34:47 <brinzhang> I dont think we should keep current logical, if so, maybe when we do next cycle to add a device, we cannot get the device occupied in obj 03:35:45 <chenke> Hi brin. "we cannot get the device occupied in obj" means what? 03:35:48 <Yumeng> xinranwang: ok, will let you know later. I will discuss with Sundar, and maybe we can create one! 03:36:03 <xinranwang> chenke: I think this patch does solve problems, it does not change the main logic, just revert the cyborg DB. I think it is useful. 03:36:25 <chenke> Agree. 03:36:28 <brinzhang> IMHO, the logic of the current code in this file makes it complicated to read. If possible, I hope to be able to refactor it. 03:36:31 <xinranwang> Yumeng: thanks！ 03:36:53 <brinzhang> xinranwang, right, it just resolve the problem, not change any logical 03:37:16 <chenke> brinzhang. Agree. If we have better solution. I want to talk about it. 03:37:42 <Yumeng> brinzhang: yes. let's discuss it more detaily. 03:37:45 <brinzhang> Yumeng: We can take https://review.opendev.org/#/q/status:open+project:openstack/cyborg+branch:master+topic:unittest.mock to PTG 03:38:03 <xinranwang> brinzhang: Yes, we are trying to find a way to decouple the code. There is still some gaps. We'd better let it be like it is(the main logic) in this release, and we can discuss in PTG. 03:38:18 <brinzhang> Sundar think we cannot do this right now, but whatever we shuold do sooner or later 03:39:12 <brinzhang> yes xinranwang, https://review.opendev.org/#/c/718584 isnot change it's logical, we can looked into later 03:39:20 <Yumeng> ok. I will paste the PTG etherpad later. let's add topics. 03:39:44 <brinzhang> https://review.opendev.org/#/c/717519/. 03:39:53 <brinzhang> https://review.opendev.org/#/c/717520/ 03:40:29 <brinzhang> above two patches, I think we should merge, because of droped py27 from Cyborg 03:40:50 <brinzhang> many projects are done of this cleanup 03:41:08 <xinranwang> Another thing: There is still some operations are not supported in nova like shelve/unshelve. I'd like to investigate it and do this part in next release if possible. I will add this topic to PTG etherpad too. 03:41:42 <Yumeng> Great! Thanks Xinran! 03:41:53 <Yumeng> will review them Brin. 03:41:59 <Yumeng> if nothing else. Let's wrap up this meeting and I'll see you all again next week 03:42:20 <Yumeng> Thank you all. have a good day! 03:42:29 <xinranwang> Thanks Yumeng 03:42:33 <xinranwang> Bye bye all 03:42:34 <Yumeng> Bye 03:42:42 <Yumeng> #endmeeting