02:01:58 #startmeeting mogan 02:01:59 Meeting started Thu Oct 12 02:01:58 2017 UTC and is due to finish in 60 minutes. The chair is zhenguo. Information about MeetBot at http://wiki.debian.org/MeetBot. 02:02:00 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 02:02:02 The meeting name has been set to 'mogan' 02:02:28 hi 02:02:41 liusheng, shaohe_feng, wanghao, litao__, Xinran, zhangyang: o/ 02:03:07 hah 02:03:32 o/ 02:05:14 o/ 02:05:40 as always, the agenda: 02:05:40 #link https://wiki.openstack.org/wiki/Meetings/Mogan#Agenda_for_next_meeting 02:05:48 let's jump in 02:05:54 #topic announcements and reminders 02:06:15 I created a library https://launchpad.net/os-cpu 02:06:26 o/ 02:06:42 Xinran: o/ 02:07:11 will move the shared codes there 02:07:21 to avoid duplication and drift with nova 02:07:31 currently it placed here https://github.com/niuzhenguo/os-cpu 02:07:44 still empty :D 02:08:00 why named os-cpu? 02:08:11 os-> openstack cpu-> compute 02:08:20 okay 02:08:23 like os-brick and os-vif 02:08:33 so Nova and Mogan will move shared codes to here? 02:08:48 only Mogan use and maintain it 02:09:26 if it's ok, hope nova can use it later 02:09:34 yeah, hope that 02:09:46 anyway, we should do that 02:09:46 that's a right way 02:09:55 okay 02:09:58 zhenguo: should we mention that in mailing list ? 02:10:11 liusheng: I think so 02:10:20 liusheng: sure, later I will 02:10:23 at least let nova guys know that 02:10:25 ok 02:10:29 after we created a repo under openstack 02:10:52 I will move configdrive/metadata stuff there first 02:10:58 to make it a library 02:11:17 zhenguo: hah, not sure if the repo should belong to mogan group or others 02:11:50 not sure 02:11:59 but I prefer not, hah 02:12:24 we can maintain it, but not as a sub project for us 02:12:34 zhenguo: ok, let's take a look of os-brick and os-vif later 02:12:43 en, hope nova guys can move codes to here too. 02:12:44 zhenguo: yes 02:13:16 wanghao: Sean said, nova use os-brick after cycles 02:13:42 zhenguo: yes 02:13:43 so we should do it first, and well define the library 02:13:59 zhenguo: morningb 02:14:09 shaohe_feng: o/ 02:15:04 ok, another thing is our official application https://review.openstack.org/#/c/508400/ 02:15:30 please keep an eye on that 02:15:46 恩 02:15:48 en 02:17:15 https://github.com/openstack/mogan 02:17:26 two +1 02:17:26 we are back on our own feet now :D 02:18:09 but seems something wrong with the gate now 02:18:59 yes, seems other projects also broken 02:19:15 yes, seems related to the zuul v3 migration 02:19:45 we can do other tings until the gate back to normal 02:20:51 not much of announcement, anyone else got a thing here 02:21:16 if not, let's move on 02:21:20 #topic task tracking 02:21:25 #links https://etherpad.openstack.org/p/MoganWhiteBoard 02:22:16 oh, seems we forget to sync with ptg discussion on the white board 02:22:46 indeed 02:23:28 please add your task there, hah 02:23:49 zhenguo: ok, sure 02:23:50 based on this https://etherpad.openstack.org/p/mogan-queens-ptg 02:23:50 should we remove the done job? 02:24:03 wanghao: yes 02:24:06 seems too much things here 02:24:08 okay 02:24:23 let's refresh it. 02:27:56 ok, anyone wants to give some updates of your task here? 02:28:59 zhenguo: I'm looking into rsd-lib 02:29:09 shaohe_feng: how 's taht 02:29:14 zhenguo: will update the spec. 02:29:33 zhenguo: send a email to yanlin for some questions, cc you. 02:29:38 zhenguo: you can check it. 02:29:43 shaohe_feng: you can add a new spec, then I will get rid of the valence one 02:29:56 zhenguo: agree. 02:29:57 shaohe_feng: ok 02:31:21 shaohe_feng: please have a look at this https://review.openstack.org/#/c/510724/ 02:31:31 not sure if client gate broken as well 02:31:50 zhenguo: OK 02:31:57 liusheng: so, how's the quota work 02:32:19 zhenguo: didn't start yet.. 02:32:55 liusheng: ok 02:32:59 zhenguo: but I am trying to see how other project's implementaiton 02:33:07 liusheng: good 02:33:49 Xinran: I find you patch about admin password support, thanks! 02:34:14 zhenguo, np 02:34:19 zhenguo: seems Keytone has plance to unify all the resource quota 02:34:26 zhenguo: * plan 02:34:51 liusheng: you mean move quota to keystone? 02:35:03 zhenguo, I am trying to return the password within the reponse of server create 02:35:06 FYI, this one: http://specs.openstack.org/openstack/keystone-specs/specs/keystone/ongoing/unified-limits.html 02:35:28 zhenguo: yes, currently, all the Cinder, Nova, Neutron have their own quota implementation 02:35:30 wxy_: thanks :D 02:36:36 liusheng: we can look into the specs 02:36:45 Xinran: cool 02:36:51 zhenguo: the implementations are not very suitable and have some drawbacks 02:36:53 zhenguo: ok 02:38:00 litao__: are you around? 02:38:10 yeah. It's block now. The first thing is to approve this two specs: https://review.openstack.org/#/c/441203 https://review.openstack.org/#/c/455709 02:39:17 wxy_, liusheng: should we wait for that happen ? 02:39:29 I'm try to take over them now. So If the quota in mogan is not urgent. I suggest you guys can follow Keystone. 02:39:43 wxy_: ok, thanks 02:39:56 wxy_: thanks for xiyuan mentioning that 02:39:56 Maybe it can be done in R release. 02:40:05 ok 02:40:17 wxy_: so we can leverage that in S :( 02:40:50 R is OK if everything goes well. 02:40:58 I'm not sure any way. :) 02:41:03 wxy_: ok, hah 02:41:16 liusheng: please keep an eye on that 02:42:01 zhenguo: ok, sure 02:42:35 ok, the unmange servers specs need more reviews 02:42:54 wxy_: http://specs.openstack.org/openstack/keystone-specs/specs/keystone/ongoing/unified-limits.html 02:43:01 I can not open it 02:43:03 wanghao: is that only for unmanaging or including transfer? 02:43:18 shaohe_feng: you can't? 02:43:19 shaohe_feng: we can open it 02:43:37 shaohe_feng: I can't believe there's a block in your office, hah 02:45:42 shaohe_feng, I can open it 02:46:34 ok, so anything else here? 02:47:19 zhenguo: liusheng: then some thing wrong with my browser 02:47:38 shaohe_feng: seems yes 02:48:59 let's move on 02:49:04 #topic open discussion 02:49:12 who's got a thing here? 02:50:04 do you think the resource classs only scheduling is rigid? 02:50:40 zhenguo: yes, a bit, hah 02:50:55 how about move to inventory based scheduling 02:51:16 I don't mean to move back to nova's vcpu, mem, disk 02:51:23 but also move flexible 02:51:35 *more 02:51:50 zhenguo: what inventories will be reported ? 02:52:07 from ironic node properties 02:52:17 not sure yet 02:52:28 and get rid of the flavor? 02:53:08 http://lists.openstack.org/pipermail/openstack-dev/2017-September/122594.html 02:53:22 see more information above from the ironic guys 02:53:46 zhenguo: ok, will read it 02:54:35 zhenguo: not sure if we will need to care about some resources' quantity 02:55:01 liusheng: seems yes, with placement traits 02:55:57 zhenguo: maybe mogan-engine directly call placement api and get rid of scheduler, hah 02:56:14 liusheng: hah 02:56:40 liusheng: we currently lock the whole scheduleing process 02:56:53 zhenguo: yes 02:57:01 liusheng: really frustrating 02:58:11 liusheng: we can look into it to see whether we really need a scheduler service 02:58:49 zhenguo: ok, need a deep dive 02:58:54 yes 02:59:02 ok, it's almost time 02:59:12 if not other things want to discuss we can end here 02:59:31 thanks y'all! 02:59:36 #endmeeting