02:00:54 #startmeeting mogan 02:00:55 Meeting started Thu Jun 8 02:00:54 2017 UTC and is due to finish in 60 minutes. The chair is zhenguo. Information about MeetBot at http://wiki.debian.org/MeetBot. 02:00:56 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 02:00:59 The meeting name has been set to 'mogan' 02:01:40 hi all, 02:01:44 o/ 02:01:54 o/ 02:02:50 shaohe_feng, wanghao, luyao, Xinran, zhangyang: are you here? 02:02:59 o/ 02:03:08 o/ 02:03:28 ok 02:03:41 as always, the agenda: 02:03:41 #link https://wiki.openstack.org/wiki/Meetings/Mogan#Agenda_for_next_meeting 02:03:48 let's jump in 02:03:54 #topic announcements and reminders 02:04:17 we changed to use node resource class instead of adding a new node_type property to match our flavors. 02:05:21 no need to set property in ironic node? 02:05:22 but seems we can't use plancement to track node resouces now 02:05:30 litao__: yes 02:05:44 zhenguo: great 02:05:47 litao__: ironic node has a resource_class propery 02:06:06 zhenguo: is it same as node_type ? 02:06:24 liusheng: yes, we just use it as node_type 02:06:38 zhenguo: got iy 02:06:45 liusheng: and that field is just for such purpose 02:06:57 zhenguo: ok 02:07:06 as ironic also want to let placement to track resource like what we have done 02:07:26 why not mogan? 02:07:40 litao__: hah 02:07:44 hi guys, I'm here 02:07:50 wanghao: o/ 02:08:16 litao__: that's also benefet for us, we can use placement to track node resouces later 02:08:37 zhenguo: got it 02:08:50 litao: but we are a bit different, as we need to expose nodes to users to manage the node aggregates 02:09:22 if we move all our node resources to placement, users can't know which node they can add/remove to aggregates 02:09:38 that's the gap, otherwise we can just turn to Placement 02:10:43 ok, another thing is still the CLI 02:11:13 after we introduce the node notion to mogan, seems we can't avoid collisions with baremetal node 02:12:06 we recently land the patch to expose node list api to admins, but I'm still not sure how to add it to CLI 02:12:17 looks like nova 02:12:23 like openstack baremetal node list? which is what ironic use 02:12:57 lol 02:13:12 litao__: yes, but nova is host aggreagtes, so they can move all resouces tracking to placement 02:14:16 openstack bm node list seems a good alternative to me, wdyt> 02:14:23 yes, nova only exposes host to users 02:14:57 maybe openstack mogan node list hahah 02:15:13 how about we don't export "node list" command in mogan 02:15:39 wanghao: but there are some discussion before, the community disgree to use project name 02:15:56 liusheng: if so, how can admins manage node aggregates :D 02:16:37 zhenguo: admin users can just use baremetal node list of ironic, and create node aggregates through mogan 02:17:04 zhenguo: okay I see 02:17:21 liusheng: not sure if it's accepatable 02:18:06 I think use two type API is strange 02:18:20 zhenguo: for OSC perspective, we don't need to care where the baremetal node resource existing 02:18:53 zhenguo: we only need to know there is baremetal node resources in OpenStack 02:19:08 liusheng: yes, but we for mogan we only want to expose available nodes to users, not only a ironic node list 02:20:51 zhenguo: even after we exposing nodes of mogan to end user, users only can see the available nodes ? 02:20:51 and for cloudboot driver, how could we handle this? 02:21:14 zhenguo: hmm, yes, that need to consider 02:21:26 maybe admin can see all the nodes from different drivers? 02:21:30 liusheng: users just need to list the nodes through mogan, they don't need to care about the filters 02:21:44 liusheng: as we only save available nodes for mogan which can be provisioned 02:22:29 zhangyang: yes, but currently, we can't only support one driver in a deployment :D 02:23:05 s/can't/can 02:24:44 maybe we can discuss this later, also have something to discuss during the meeting 02:24:49 let's move on 02:24:49 zhenguo: yes, didn't see the need of using different drivers in a deployment now 02:25:01 ok 02:25:12 we can use different hypervisor type like nova, 02:25:26 zhangyang: yes 02:25:40 litao__: but we only have one engine service, not multi compute service 02:25:56 multibackend support need more discuss 02:26:08 sure 02:26:11 we need some feedback from really user. 02:26:31 so we can leave it as it is now. 02:26:46 yes, but for now, seems not requirement for multi backend 02:26:55 sure 02:27:32 the last thing I want to mention is the reviews 02:27:39 Contribution into mogan for the last 30 days 02:27:39 #link http://stackalytics.com/report/contribution/mogan/30 02:28:31 the last 30 days we only got 139 reviews 02:28:44 only got 1.4 per core per day 02:29:01 :( 02:29:36 :( 02:30:04 hah, just want to get more reviews from you guys, and also we want got get more core reviewers 02:30:38 lol 02:32:06 litao__: glad to see you reviewing my patches :D 02:32:31 zhenguo: my pleasure 02:32:35 sorry absent a while 02:32:41 hah 02:32:52 we really need to improve reviewing 02:33:13 liusheng: yes we do 02:34:09 Recently, I found some foreigners committed codes 02:34:17 we can slow down a bit the development. 02:34:21 zhenguo: morning 02:34:30 litao__: yes, but just some trival fix 02:34:33 shaohe_feng: o/ 02:35:09 zhenguo; yes, but a big progress 02:35:19 litao__: hah 02:35:56 ok, not much of announcement, anyone else got a thing here? 02:36:15 Do we refresh our task? 02:36:26 any update this week? 02:36:29 wanghao: yes, during next topic 02:36:34 ok 02:36:43 ok, let's move on 02:36:48 #topic task tracking 02:36:59 #links https://etherpad.openstack.org/p/MoganWhiteBoard 02:37:06 we can go through it one by one 02:37:58 first is node aggregates, I made a spec for it, need more reviews 02:38:30 and the patch to expose nodes to admins is already landed. 02:39:04 cool 02:39:10 spec and code we both have 02:39:59 wanghao: hah, just the first patch :D 02:40:08 zhenguo: good start 02:40:11 haha 02:40:18 lol 02:40:20 cool, we still need a little time to finish the aggregate part.. 02:40:56 zhangyang: hope you and liudong can review the spec to see if it meets your requriements 02:41:04 excited 02:41:46 zhenguo: liudong added some comment 02:41:56 zhangyang: ok, will check 02:42:32 ok, next is server groups and scheduler hints 02:43:11 liusheng: it's yours 02:43:52 zhenguo: ok, I have uploaded two paches for scheduler hints 02:44:12 https://review.openstack.org/#/c/463534/ 02:44:26 liusheng: yes, I see, seems almost good 02:44:52 firstly, the patch expose the api that allow users specifying scheduler_hints when creating servers 02:45:09 hi all, please remember to update the ehterpad, we plan to send out the update every week to present to the whole community 02:45:26 and the scheduler hints will be put in to request_spec 02:45:43 and will be passed to scheduler 02:46:15 liusheng: yes, and the server group patch, maybe you also need to make more time on it, as liudong seems cant get time, but he can help to review 02:47:13 zhenguo: yeah, sure 02:47:25 liusheng: ok, thanks 02:47:49 next is adopt servers, wanghao^^ 02:47:54 okay 02:48:08 this week I updated the spec according to the comments 02:48:25 aha, yes, we will try to do more reviews on this 02:48:40 yeah 02:48:47 I think we are close 02:49:11 wanghao: I also think so, thanks for sticking on this 02:49:11 so maybe next week we can prepare to code something. 02:49:31 sounds cool :D 02:49:31 zhenguo: haha happy to 02:49:53 next is valence integration 02:50:10 not sure if we should place it in high priority section 02:50:15 shaohe_feng: wdyt? 02:50:42 zhenguo: which task 02:50:51 shaohe_feng, luyao, Xinran: I assign it to us 02:51:06 zhenguo: I have seen it. 02:51:08 shaohe_feng: please look at this https://etherpad.openstack.org/p/MoganWhiteBoard 02:51:14 the valence related one 02:51:36 zhenguo: Let me check with yanglin 02:51:39 first 02:52:00 I should know the valence status. 02:52:04 shaohe_feng: ok, so we can leave a not updates to the task this week 02:52:18 OK. 02:52:24 shaohe_feng: we will sent the task report out, so valence guys can also see it 02:52:44 Great. 02:52:57 ok, seems not much time left 02:53:19 every ower can upade the other tasks on the etherpad, 02:53:29 OK BTW I move specifying port id patch into this etherpad. 02:53:41 wanghao: cool, thanks 02:53:58 wanghao: on my tests, it's almost good 02:54:13 zhenguo: thanks 02:54:27 zhenguo: I'm building my new devstack for mogan 02:54:49 wanghao: any bad thing happened? 02:54:57 zhenguo: not yet for now 02:55:05 wanghao: hah, good luck 02:55:12 zhenguo: that will help to test mogan more. 02:55:18 zhenguo: haha 02:55:25 wanghao: lol 02:55:57 wanghao: we can discuss later about the weekly report context, maybe just like ironic's 02:56:11 zhenguo: sure 02:56:38 ok, let's move on to next topic although there's only 5min left 02:56:44 #topic open discussion 02:56:50 who's got a thing here 02:57:26 no from me. 02:58:09 hah, ok, if not, we can end the meeting 02:58:20 thanks y'all 02:58:23 cool bye bye 02:58:26 bye 02:58:36 #endmeeting