13:01:12 <joehuang> #startmeeting tricircle 13:01:12 <openstack> Meeting started Wed Aug 31 13:01:12 2016 UTC and is due to finish in 60 minutes. The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:01:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:01:15 <openstack> The meeting name has been set to 'tricircle' 13:01:19 <zhiyuan> hello 13:01:21 <yinxiulin> hello 13:02:14 <joehuang> #topic rollcall 13:02:21 <joehuang> #info joehuang 13:02:33 <yinxiulin> #info xiulin 13:02:34 <dongfeng> #info dongfeng 13:02:34 <Yipei> #info Yipei 13:03:20 <Rong_hui> :) 13:03:26 <zhiyuan> #info zhiyuan 13:03:38 <Rong_hui> #info ronghui 13:03:43 <joehuang> #topic feature implementation progress review 13:03:43 <hejiawei> #info hejiawei 13:04:06 <joehuang> hello, please describe your feature implementation progress shortly 13:05:07 <Yipei> i have already submitted a patch, i will update the patch based on your comments. Further, i will try to finish the patch of adding resource_affinity_tag ASAP 13:05:35 <Yipei> go on updating the patch of dynamic pod binding 13:05:46 <joehuang> to Yipei, except resource_affinity_tag, also one field in pod binding 13:05:57 <joehuang> to represent current active binding 13:06:19 <dongfeng> I have commit a document of pod api features. tempest test for pod is still going on. 13:06:34 <zhiyuan> patches for resource clearing of floating ip, subnet, router have been submitted, but they all depends on the shared-vlan-l3 patch, which has not been merged 13:06:51 <joehuang> good, to Dongfeng, do you know how to do the tempest test? 13:07:07 <yinxiulin> I'm doing server action function. The action which only need server_id parameter has been finished, and passed the unit test. 13:07:24 <dongfeng> no, tempest doesn't contain the pod test cases. i am studying it. 13:07:32 <zhiyuan> now i am working on the separation of api-gateway and networking automation 13:07:35 <joehuang> to Zhiyuan, so we need to merge shared-vlan-l3 first, so that other patches can get merged 13:07:38 <Yipei> do we need to add a table to store historical binding relationship, instead of adding a field in binding table? i mean all the records in the binding table are active 13:07:56 <zhiyuan> to joe, that's right 13:08:02 <hejiawei> I read some document to learn how to test my patch ,and next week I will continue to test it . 13:08:03 <joehuang> to Xiulin, great! 13:08:04 <Yipei> or just add a field 13:09:43 <joehuang> no need to add a new table, just use active, deactive, and the time related field for example update_at to store the history information 13:10:20 <joehuang> to Zhiyuan, good job, quick reaction to new proposal 13:10:56 <joehuang> to Jiawei, learn from other's patch will accelearate your experience on how to do test 13:11:07 <Yipei> to joe, ok, got it. will add it into the binding table 13:11:13 <Rong_hui> For me, the environment meet some problems, and i solve it and record it in the our test and uploading to wiki 13:12:11 <Rong_hui> and also do some creating the new organization for our team. 13:12:36 <zhiyuan> to ronghui, is the installation guide in wiki page ready? 13:13:15 <Rong_hui> i will finish it this week and add some case for it 13:13:24 <joehuang> I suggest you guys who has no lots of experience on test, need to learn unit test and tempest test 13:14:13 <Rong_hui> ok 13:14:16 <dongfeng> ok 13:14:23 <Yipei> ok, got it 13:14:33 <hejiawei> I got it 13:14:40 <longxiongqiu> ＃info longxiongqiu 13:14:40 <zhiyuan> document for mock: http://www.voidspace.org.uk/python/mock/ 13:14:53 <longxiongqiu> hi joehuang, i will update my patch ASAP 13:15:12 <longxiongqiu> and continue my working ASAP 13:15:14 <Rong_hui> xiongqiu has come back to our team 13:16:05 <joehuang> hi, Xiongqiu, shake hand 13:16:41 <longxiongqiu> ：） 13:16:43 <joehuang> long time no see 13:17:49 <longxiongqiu> yes, i just come back yesterday 13:19:04 <joehuang> I'll try to find some material about unit test and functional test. and share that to you in #openstack---\\-\\\-tricircle 13:19:19 <joehuang> #openstack-tricircle 13:19:29 <Rong_hui> ok i think is very helpful to us 13:20:17 <joehuang> #topic proposal to move tricircle forward for its big-tent application 13:20:23 <Rong_hui> also some material about the tempest 13:21:28 <joehuang> hello, last week we talked about that the TCs concerns on Tricircle big-tent application 13:21:43 <joehuang> and the proposal to move forward was prepared 13:21:50 <joehuang> #link https://docs.google.com/presentation/d/1kpVo5rsL6p_rq9TvkuczjommJSsisDiKJiurbhaQg7E/ 13:22:53 <joehuang> #info the basic idea for the proposal 2 is to divide tricircle into two sub-projects 13:23:19 <joehuang> #info one for networking automation 13:24:02 <joehuang> #info the other one for API Gateway for Nova/Cinder 13:24:21 <joehuang> so would like to know your comments 13:25:12 <joehuang> For proposal, plugin mechanism in Nova/Cinder will take long time to discuss 13:26:37 <joehuang> the proposal 2 will also make tricircle being more flexible in deployment for different scenario 13:28:28 <zhiyuan> the difference between proposal 2 and our current implementation is that we also need to add a new plugin in bottom Neutron server 13:28:28 <Rong_hui> i dont really understand about the plugin mechanism 13:29:41 <joehuang> to Ronghui, you mean Neutron or Nova/Cinder? 13:30:02 <Rong_hui> Cinder 13:30:03 <joehuang> Nova/Cinder don't support plugin mechnism yet in its API layer 13:30:34 <Rong_hui> ok 13:30:35 <liukun> #info liukun 13:31:07 <joehuang> For Nova/Cinder, if we want to add plugin mechanism then we have to develop that in Nova/Cinder, and need Nova/Cinder to approve it 13:31:18 <joehuang> hi liukun, welcome back 13:31:29 <liukun> thx 13:31:43 <joehuang> need long time discussion in Nova/Cinder team 13:31:53 <Rong_hui> got it 13:32:19 <Yipei> why do nova and cinder need plugins, in current implementation, we do not use plugins 13:32:39 <Rong_hui> +1 13:33:32 <joehuang> there is also proposal for Nova API-GW and Ciner API-GW for global objects like volume_type, flavor implementation to remove the overlapping implementation with Nova/Cinder 13:34:20 <shiyangkai> #info shiyangkai 13:34:37 <zhiyuan> in our current implementation, we need to implement nova/cinder api gateway, we need to handle api request ourselves, parse and validate the request body, process and return the response body 13:35:21 <joehuang> to Yipei, Because TCs worried about that the tricircle will introduce different in the Nova/Cinder API 13:35:23 <zhiyuan> if plugin mechanism is introduced, we can focus on the processing, no need to care about the request and response body 13:36:14 <joehuang> TCs want single API implementation and always keep the API consistency with the openstack documentation 13:36:21 <Yipei> ok, i see. actually, we do not introduce any different API 13:36:40 <joehuang> syntax or sematics difference 13:37:44 <joehuang> we have already used tempest to guarantee the syntax consistency, but they still worried that sematics difference will be there someday 13:38:09 <joehuang> welcome shiyangkai 13:38:51 <joehuang> how do you think about proposal for Nova API-GW and Ciner API-GW for global objects like volume_type, flavor implementation 13:38:54 <joehuang> to remove the overlapping implementation with Nova/Cinder 13:39:02 <shiyangkai> thank you very much 13:40:30 <zhiyuan> +1 for this proposal 13:40:35 <joehuang> before project dividen( we also look forward response from community and Monty/TCs etc 13:41:22 <joehuang> we can do first to move the networking automation trigger from Nova API-GW to local neutron plugin 13:41:51 <joehuang> and optimize the volume_type/flavor etc implementation 13:42:18 <joehuang> no matter we divide the project or not, these two proposal brings benefit to Tricircle 13:42:55 <joehuang> we can wait for comment from community to see if we need to divide the project or not 13:43:13 <Rong_hui> ok 13:43:15 <joehuang> how about your thoughts? 13:43:18 <Rong_hui> i got it 13:44:19 <Yipei> no other questions, i think it is a good idea 13:44:26 <Rong_hui> It is another way for us to let TCs accept more easy and also benefit to Tricircle 13:44:49 <joehuang> good 13:45:21 <joehuang> #info optimize tricircle to be prepared for divide 13:46:05 <joehuang> other topics? 13:46:22 <Rong_hui> no for me 13:46:39 <liukun> no 13:46:47 <yinxiulin> no for me 13:46:49 <zhiyuan> no for me 13:46:51 <Yipei> no for me 13:46:51 <dongfeng> no for me 13:47:23 <shiyangkai> no 13:47:40 <joehuang> ok, let's conclude the meeting 13:47:54 <joehuang> thank you attending the meeting, see you next time 13:47:59 <joehuang> #endmeeting