Earlier  
Posted Nick Remark
#openstack-cyborg - 2020-11-19
03:05:07 swp20 hi Yumeng
03:05:33 Yumeng hi wenping
03:06:02 brinzhang_ hi Yumeng, swp20
03:06:20 Yumeng hi Brin
03:06:24 swp20 i see the nova cores review, we have to modify a lot.:(
03:06:35 Yumeng yep.
03:06:36 Yumeng https://review.opendev.org/#/c/750116/
03:07:35 Yumeng Seems cyborg is expected to create mdev to the sys by itself
03:08:18 swp20 we can create mdev when binding arq by patch api.
03:08:55 Yumeng Yes, or even we can create when we report attach_handle
03:09:30 Yumeng but seems this may do redundant creation
03:09:52 swp20 why
03:13:01 Yumeng because everytime we change the vgpu_type eg.change from nvidia-223(num=8) to nvidia-226(num=4), we need to delete redundant attach_handles
03:13:37 swp20 hi xinranwang
03:13:38 Yumeng when we report, we need to create again.
03:13:58 xinranwang Hi all, sorry for late
03:14:47 Yumeng hi xinranwang:no pl.
03:16:50 brinzhang_ is there any others?
03:17:21 Yumeng yes? still here
03:17:31 xinranwang I am trying to catch up with you guys.
03:17:36 brinzhang_ Yumeng, xinranwang, swp20:I mean, we can start ^^
03:17:48 Yumeng sure:)
03:18:15 Yumeng #startmeeting openstack-cyborg
03:18:15 openstack Yumeng: Error: Can't start another meeting, one is in progress. Use #endmeeting first.
03:18:27 Yumeng #endmeeting
03:18:29 openstack Meeting ended Thu Nov 19 03:18:27 2020 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)
03:18:30 openstack Minutes: http://eavesdrop.openstack.org/meetings/openstack_cyborg/2020/openstack_cyborg.2020-11-12-03.07.html
03:18:31 openstack Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_cyborg/2020/openstack_cyborg.2020-11-12-03.07.txt
03:18:32 openstack Log: http://eavesdrop.openstack.org/meetings/openstack_cyborg/2020/openstack_cyborg.2020-11-12-03.07.log.html
03:18:42 Yumeng #startmeeting openstack-cyborg
03:18:43 openstack Meeting started Thu Nov 19 03:18:42 2020 UTC and is due to finish in 60 minutes. The chair is Yumeng. Information about MeetBot at http://wiki.debian.org/MeetBot.
03:18:44 openstack Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
03:18:46 openstack The meeting name has been set to 'openstack_cyborg'
03:19:12 brinzhang_ This endmeeting is privos meeting, xinwanwang send endmeeting, but it's not useful
03:19:18 Yumeng #topic Roll call
03:19:29 brinzhang_ #info brinzhang_
03:19:37 xinranwang Ops, it did not work..
03:19:42 xinranwang #info xinranwang
03:19:47 swp20 #info swp20
03:20:00 brinzhang_ yes, it's must run by the meeting start owner
03:20:02 Yumeng maybe it needs the same host..
03:20:09 Yumeng =_=!
03:20:26 Yumeng yes, the same person to do the operations
03:20:38 xinranwang lol nice emoji
03:20:39 brinzhang_ right~
03:21:03 Yumeng lol
03:21:06 swp20 =_=
03:21:45 Yumeng https://wiki.openstack.org/wiki/Meetings/CyborgTeamMeeting#Weekly_IRC_Cyborg_team_meeting
03:24:07 swp20 Yumeng: what do you means by report attach_handle, to Placement or Nova?
03:27:33 Yumeng I mean when agent generates the attach_handle
03:29:48 swp20 if we change vgpu type, we ensure to delete the old attach handle.
03:30:06 Yumeng yes, that's what I mean.
03:30:19 Yumeng old and unused attach_handle
03:31:20 Yumeng so probably create mdev during arq bind is a better choice
03:31:46 swp20 if the attach handle have bound with instance, can we delete it?
03:31:53 Yumeng anyway, I will update nova spec and fix their comments.
03:32:03 swp20 cool
03:33:07 Yumeng we cannot delete a single attach_handle. but we can delete a mdev_UUID when unbound arq
03:33:56 swp20 ok got it.
03:34:01 Yumeng cool
03:34:22 Yumeng # topic smartnic
03:35:44 Yumeng cyborg side spec needs review :https://review.opendev.org/#/c/759545/3
03:37:45 Yumeng and program API improvement feature. is anyone intrested in this topic>?
03:38:05 xinranwang yes, please review it. hope we can get this merged first.
03:38:28 Yumeng that's one of our wallaby goal: https://wiki.openstack.org/wiki/Cyborg/Wallaby_Release_Schedule
03:38:45 swp20 LGTM
03:39:49 xinranwang swp20: thanks
03:40:30 Yumeng xinranwang: will review. :)
03:40:41 xinranwang Yumeng: thanks!
03:40:45 Yumeng annd also cyborg-side vGPU driver https://review.opendev.org/#/c/758925/
03:41:02 brinzhang_ smartnic spec I will review today ^
03:41:24 Yumeng annd also cyborg-side vGPU driver https://review.opendev.org/#/c/758925/ pls review it.
03:41:32 brinzhang_ Yumeng: vGPU driver spec, you seems not completed the comments from me, I left comment
03:41:56 brinzhang_ completed that, I am ok
03:42:12 Yumeng ha? which one ?I think I've fixed all the comments.
03:42:33 brinzhang_ left comment in the spec
03:43:07 xinranwang I will review vgpu today
03:43:17 brinzhang_ Yumeng, xinranwang, swp20: we have many open status patch need to review, if you have time please review https://review.opendev.org/#/q/project:openstack/cyborg+status:open
03:44:06 xinranwang ok
03:44:49 Yumeng I am confused by the comment"Where is the NOTE?"
03:44:55 Yumeng brinzhang_
03:45:48 brinzhang_ I would like you can explain the nova side, just have a simple is ok, IMO
03:46:07 brinzhang_ .. note: that why I want to say
03:46:25 brinzhang_ s/why/what
03:47:46 Yumeng do you mean "needs to mention OWNER_NOVA " here?
03:48:18 brinzhang_ when we back review this spec after several releases, we can know that change will be contained nova and cyborg
03:48:26 brinzhang_ yes
03:48:56 Yumeng brinzhang_: sure. that's not a problem. Will add it.
03:49:07 brinzhang_ thanks
03:49:49 Yumeng I actually already mentioned that in nova side spec and we've already included the nova spec in this one.
03:51:03 brinzhang_ Yumeng: I think it's useful to point out some key information
03:51:12 Yumeng anyway, will update
03:52:11 Yumeng Do you guys have any other things to bring up for today?
03:52:27 brinzhang_ nothing
03:52:37 xinranwang nothing from me
03:53:33 Yumeng ok~ then see you guys.
03:53:43 Yumeng Lunch time
03:53:55 Yumeng Have a good day~ Bye
03:54:02 xinranwang bye bye
03:54:10 Yumeng #endmeeting

Earlier   Later