13:02:12 #startmeeting tricircle 13:02:13 Meeting started Wed Aug 24 13:02:12 2016 UTC and is due to finish in 60 minutes. The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:02:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:02:16 The meeting name has been set to 'tricircle' 13:02:33 #topic rollcall 13:02:37 #info joehuang 13:02:51 #info Yipei 13:02:52 #info dongfeng 13:02:57 #info xiulin 13:03:02 #info zhiyuan 13:03:53 #topic feature implementation progress 13:04:29 hi, please share your recent progress in feature implementation, thanks 13:05:56 patches for subnet and floating ip clearing have been submitted, i am going to submit the patch for router clearing tomorrow 13:05:59 Zhiyuan fixed one py27 test error today, so please rebase your patch to pass the py27 check and gate test 13:06:30 I resubmitted a patch based APP implementations for microversion 13:06:48 thanks to zhiyuan, this is necessary for nova action to pass the tempest test 13:06:48 about dynamic pod binding, the unit tests fail locally. so spends a lot time on tests. but i just submitted a patch, all the tests pass. 13:07:06 this week i will revise commit message and submit a patch. i will add you guys in the reviewer list 13:07:22 I have finished the documentation on pod and pod binding. but they are not merged yet. 13:07:30 that means all unit test passed for dynamic pod binding? 13:08:03 to Dongfeng, have you tried the command for pod and pod binding? 13:08:20 to Xiulin, great! 13:08:47 this week,I commit a patch that force a volume to detach,but not passed .this is my first time to commit,I will retest it and thanks all the reviewers 13:09:01 yes, after the releveant functions by curl, i look up the response code in screen -x 13:09:05 FOr me, the policy patch has been updated according to the review comment 13:09:40 to Jiawei, thank you very much for the first commit! 13:11:01 after i run the command by curl, i look up the response code in screen -x. it could be verified that the response code is 200 if correct. 13:11:22 if you found your patch can't pass the py27 test, please rebase your patch 13:11:44 to Dongfeng, yes, it should be like that 13:12:55 any more information on feature implementation? 13:13:08 ok, i will try. i did not add more tests, just try to checks whether need to dynamically bind a pod when creating a vm. 13:13:18 everybody is encouraged to review other's patch 13:13:52 i will submit a patch without [WIP] tag this week, and add you guys in the reviewer list. 13:14:15 Thank you very much, Yipei, go ahead 13:15:01 ok, let's talk about the TC concerns on tricircle big tent application 13:15:12 you are welcome, it is my duty 13:15:30 #topic TC concerns on tricircle big tent application 13:16:00 Tricircle has started to apply big tent project last month 13:16:29 and two rounds dicussion in the TC weekly meeting 13:17:11 a lot of review comment in the patch 13:17:47 #link https://review.openstack.org/#/c/338796/ 13:19:22 #info the main concerns are API consistency in Nova API-GW / Nova, Cinder API-GW/Cinder 13:20:32 #info but the value of networking automation across Neutron is recogized 13:21:42 what's your idea to move tricircle forward? 13:22:00 or proposal 13:23:05 so it's better if our networking automation features can work with classical Nova and Cinder server 13:23:44 you mean decouple the networking from Nova-API-GW and Cinder API-GW? 13:24:32 Currenly Nova API GW is involved in networking automation 13:24:42 yes, so the main concern is away :) 13:25:31 that means let Nova/Cinder still work in multi-region mode 13:26:40 #info idea1, to make networking automaton work with multi-region Nova/Cinder, decouple from Nova API-GW 13:26:58 any other idea or proposal? 13:27:50 samueldmq: it uses an extra random number as part of the key so that when we change it (invalidate) the keys are effectively invalidated 13:28:48 this is a good idea, no need to worry about Nova API-GW/Cinder API-GW will have overlapping/different API implementation 13:29:14 sorry i am late.I have watched sound code this week and prepared to code in the next few days. 13:29:53 I also have another proposal to introduce plugin mechanism in Nova Cinder API layer 13:30:33 but it'll take long time to disuss in Nova/Cinder to support such a feature 13:31:01 for decoupling, it's not that straightforward since Nova server doesn't provide driver framework, you can only write driver for nova compute 13:31:14 #info idea2 propose to introduce plugin mechanism in Nova Cinder API layer 13:31:37 agree 13:31:53 but we need VM booting event to trigger networking automation 13:32:47 yes 13:33:31 so, zhiyuan, could you draft the idea1, and let's have more and broad discussion on this topic to see if it's possible 13:34:38 ok, i will prepare for the discussion 13:35:05 but for our current development, just like what TC expect, continue our current implementation 13:35:19 don't stop it 13:35:55 #action discussion on the possiblity to decouple the networking automation 13:35:56 ok, got it 13:36:51 ok 13:36:59 good, please continue our development and think about how to address TCs' concerns and worries 13:37:18 #topic open discussion 13:37:37 any other topic? 13:38:47 no from me 13:38:57 no for me 13:39:02 no for me 13:39:07 no for me 13:39:34 ok, thank you very much, let's conclude the meeting 13:39:43 #endmeeting