02:00:45 #startmeeting mogan 02:00:46 Meeting started Thu Mar 23 02:00:45 2017 UTC and is due to finish in 60 minutes. The chair is zhenguo. Information about MeetBot at http://wiki.debian.org/MeetBot. 02:00:47 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 02:00:49 The meeting name has been set to 'mogan' 02:01:22 here 02:01:32 o/ 02:03:19 morning 02:03:46 I'm herer 02:03:56 ok, morning all, 02:04:07 as always, the agenda: 02:04:07 #link https://wiki.openstack.org/wiki/Meetings/Mogan#Agenda_for_next_meeting 02:04:25 let's jump in 02:04:33 #topic announcements and reminders 02:05:06 You may have noticed the core nomination email, please respond your support or any concerns , thanks! 02:05:24 we really need more cores :P 02:06:00 Maybe i miss that email 02:06:12 I don't like to merge patches directly by myself, please do more reviews 02:06:29 litao: you can get it if someone else respond it:D 02:07:03 zhenguo: ok’ 02:07:15 hah, liusheng is absent today 02:08:02 o/ 02:08:07 The refactors are almost finished, we have moved scheduler out and save sompute nodes/ports resources to DB instead of cache. 02:08:16 RuiChen: o/ 02:09:09 liudong, zhangyang: is the base driver ok for you now? 02:09:28 sounds good, thank you zhenguo 02:09:38 hah 02:09:45 zhenguo: yes, it's great 02:10:25 ok, thanks. really many refactors recently, I'm sure there should be many following up small issues, please feel free to fix them. 02:11:17 wrt the new flavor, I would like to align with Valence guys first to see what informations needed to compose a node 02:13:04 lin_yang, ntpttr_: hey, are you around? 02:13:55 ok, seems no valence guys here, 02:13:57 let's move on 02:14:17 not much of announcement, anyone else have a thing here? 02:14:34 anyone found a bug, please commit a launchpad bug, if you don't have time to fix, other guy might contiue to work from launchpad 02:14:52 yes, thanks RuiChen for the reminder 02:15:01 sounds good 02:15:40 ok, let's move on 02:15:46 #topic task tracking 02:15:56 #link https://etherpad.openstack.org/p/mogan-task 02:16:09 * zhenguo gives folks a few minutes to review 02:17:47 looks like quotas and faults tasks moving along nicely, thanks wanghao and luyao 02:19:03 thank you wanghao and luyao 02:19:12 and sorry for not updating the task list recently due to refacotors 02:20:33 I will prepare a priority task list by next meeing, then we can follow that 02:23:06 shaohe_feng: o/ 02:23:53 zhenguo: hi 02:23:58 zhenguo: I'm here 02:24:16 morning shaohe_feng 02:24:38 zhenguo: sorry for trafic jam 02:24:42 shaohe_feng: we are talking about the tasks now, do you have anything to tell 02:24:53 RuiChen: morning . 02:25:10 zhenguo: I need to know some context. :) 02:25:16 would like to see more folks schedule themselves review queue according to the task priority 02:25:21 shaohe_feng: hah, ok 02:25:27 zhenguo: Maybe I can talk with you after meeting. 02:25:39 shaohe_feng: sure 02:25:52 hi guys, do you have anything to tell here? 02:25:59 we are always around in mogan channel 02:26:01 :-) 02:26:09 lo 02:26:10 lol 02:26:56 litao: hi, do you want to improve the multi instance creation? 02:27:20 yes, maybe need to fix something 02:27:26 litao: maybe we need to schedule for all instances first and run the workflow 02:28:14 litao: like if there's not enough nodes, you have already created some instances regardless of the min_count 02:28:20 but the scheduler process is in the workflow 02:28:35 litao: I may found an issue in multi instance creation, we can talk it later. 02:28:46 litao: I'm thinking to move it out 02:29:16 we can just leave the OnFailureRescheduleTask in the workflow maybe 02:29:35 ok, that 's fine, I will optimize it 02:29:42 thanks 02:29:58 wanghao: ok 02:31:13 I will focus on configdrive task later 02:31:14 did anyone work for moving scheduler out first? 02:31:25 litao: maybe I can 02:31:32 zhenguo: 02:31:33 good 02:32:03 litao: in fact, it's part of the whole refactors, lol 02:32:24 and the last thing is 'flavor' 02:32:28 zhenguo: why we are moving shceduler out? 02:33:00 shaohe_feng: in fact it just has a execute method now 02:33:13 zhenguo: we split configdrive? 02:33:16 zhenguo: fine 02:33:54 zhenguo: sorry, so buzy these days, no energy to follow all of you. 02:33:56 shaohe_feng: yes, I would like to land password configdrive first 02:34:16 zhenguo: greate, let implement it one by one. 02:34:31 shaohe_feng: hah, you can try to setup a devstack env, it changed most 02:34:34 zhenguo: maybe this is a easy way to push a big feature. 02:34:58 shaohe_feng:yes, and we will not support metadata way 02:35:13 zhenguo: oh, hope no painful to re-setup the devstack. 02:35:22 shaohe_feng: hah 02:35:44 zhenguo: OK. so we also not support metadata server, right? 02:36:16 shaohe_feng: yes, 02:36:30 shaohe_feng: as we can't support it :( 02:36:51 zhenguo: why? we need neutron proxy? 02:37:13 shaohe_feng:yes, and the we will conflict with nova metadata server 02:37:19 zhenguo: OK. 02:37:39 shaohe_feng: but configdrive can meet our requirements as well 02:37:47 zhenguo: then should we support new keypairs API? 02:38:04 zhenguo: OK. just support configdrive 02:38:41 shaohe_feng: configdrive is a way to store things like password, keypairs, files 02:39:22 shaohe_feng: keypair is a must for us 02:39:40 zhenguo: yes, we need to store the public key of users. 02:40:09 zhenguo: so firstly RAS key. then x509, right? 02:40:22 shaohe_feng: sounds ok 02:41:32 not sure if you have read this https://review.openstack.org/#/c/408151/ 02:42:14 zhenguo: not yet. 02:42:36 zhenguo: you have mentioned it. 02:42:44 zhenguo: BTW, should we support list servers by tags? 02:42:49 and this 02:42:49 #link https://review.openstack.org/#/c/442387/ 02:43:02 litao: sure, we should 02:43:45 zhenguo: I just seed the result, two -1, and workflow -1 02:44:01 zhenguo: so mogan should support instance RAID configuration 02:44:32 shaohe_feng: yes, the author wants us to prepare a spec for RAID configuration as well. 02:44:59 zhenguo: for mogan comes out for baremetal, right. 02:45:11 zhenguo: mogan will dedicate for baremetal. 02:45:26 maybe they want to see how we handle that 02:45:57 for RAID, I think we need to save disks info 02:46:12 zhenguo: why you not ask the author to prepare a spec for mogan? 02:46:22 zhenguo: you can invite him. :) 02:46:48 zhenguo: and we need more contributors. :) 02:46:50 shaohe_feng: he is a nova guy, hah 02:47:09 zhenguo: if he want want to use a RAID? 02:48:24 if anyone interested in this, please prepare a spec or I can do it later 02:48:42 ok, anything else here? 02:48:53 zhenguo: oh, I know, he does not care the users. He just want to write code, a "Ma nong" in Mandarin. 02:49:22 zhenguo: no for me. 02:49:30 hah, 02:49:32 let's move on 02:49:39 #topic open discussion 02:49:43 Ma nong , hah 02:49:57 who's got a thing :) 02:50:42 oh, I have fixed this issue in quota driver(part 2), that we can move on. 02:50:59 wanghao: oh, thanks 02:52:13 hi guys, do you think we need to change the meeting time to get more people get involved? 02:53:13 zhenguo: sounds good. not suer how many people will join us, and where they come from 02:53:38 hah, will check with others in our channel later 02:53:51 good 02:53:53 seems it's ok for all of us in the meeting now 02:53:59 we can send a mail in openstack-dev for this 02:54:04 sure 02:54:32 but as we have moved to openstack-meeting, not sure if the time slots is good fit for us 02:54:34 if some 'lao wai' want to join us, he can select a time. 02:54:50 litao: IMHO, a really good engineer need to consider which is the way to implement user's requirement. That's the differance between "Ma Nong" :) 02:55:01 if not, we can keep this 02:55:07 zhenguo: yes 02:55:32 ok, anything else? 02:55:45 no 02:55:45 shaohe_feng: yes 02:56:16 ok, it's almost time 02:56:24 thanks y'all 02:56:45 #endmeeting