13:02:48 <joehuang> #startmeeting tricircle
13:02:48 <openstack> Meeting started Wed Jun 15 13:02:48 2016 UTC and is due to finish in 60 minutes.  The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:02:49 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:02:51 <openstack> The meeting name has been set to 'tricircle'
13:03:03 <joehuang> #topic rollcall
13:03:13 <joehuang> #info joehuang
13:03:17 <longxiongqiu> hi
13:03:23 <longxiongqiu> #info longxiongqiu
13:03:31 <Yipei_> #info Yipei_
13:03:32 <ronghui_1> #info ronghui
13:03:38 <zhiyuan> #info zhiyuan
13:03:39 <Yipei_> hi, all
13:03:51 <joehuang> hi
13:03:51 <ronghui_1> hi
13:04:01 <joehuang> #topic tempest
13:04:33 <joehuang> a patch has been submitted to project-config
13:05:01 <joehuang> so the integration test job could be configured in the CI pipeline
13:05:25 <joehuang> you will see a new job for each patch submitted except py27 pep8
13:05:46 <zhiyuan> what test our tempest test covers currently?
13:06:02 <joehuang> after the job could be executed, we can add the test cases from tempest case by case
13:06:30 <zhiyuan> so now it's only a dummy job?
13:06:35 <joehuang> currently two type of cases test_volume_get* and test_volume_list*
13:06:44 <zhiyuan> oh i see
13:06:54 <joehuang> but the scripts need to be verified after the job can run
13:07:43 <joehuang> I checked locally, still some bugs in the post_test_hook.sh
13:07:56 <joehuang> after the scripts work
13:08:32 <joehuang> for xiongqiu's patch like cinder volume type, can use cinder tempest test cases to test the code work or not
13:09:26 <longxiongqiu> do i need to write the tempest code for my patch?
13:10:07 <joehuang> and will list all tempest test cases for nova/cinder/neutron in the scripts but commented, after a new feature added, the corresponding test cases will be un-comment and be executed to test the patch submitted
13:10:48 <joehuang> no need, just un-comment the corresponding test cases
13:11:26 <joehuang> then it should be executed in the dvsm_funcational test job
13:11:32 <longxiongqiu> ok
13:11:53 <joehuang> #info will list all tempest test cases for nova/cinder/neutron in the scripts but commented, after a new feature added, the corresponding test cases will be un-comment and be executed to test the patch submitted
13:12:34 <joehuang> currently it took long time to test your patch after you submitted
13:12:58 <joehuang> and can not make sure your code work or not
13:13:49 <joehuang> any question?
13:13:56 <zhiyuan> no
13:14:36 <longxiongqiu> but i tested it in my real env, it can work.
13:14:40 <longxiongqiu> the az_fix patch py27 test failed, have anybody know the reason?
13:15:00 <Yipei_> i have the same problem
13:15:07 <joehuang> understand, xiongqiu
13:16:03 <Yipei_> unit test in py27 fails
13:16:20 <longxiongqiu> the az_fix patch test failed in network test part, it is strange, i just changed some cinder volume creation code,not effect network.
13:17:00 <zhiyuan> you can check the log to see why
13:17:17 <joehuang> the log is : Exception: Neutron core_plugin not configured!
13:17:36 <joehuang> may be neutron changed something recently
13:17:45 <longxiongqiu> i have checked, but i don't understand it
13:18:39 <joehuang> #link http://logs.openstack.org/24/306224/14/check/gate-tricircle-python27/b0e91a7/console.html
13:19:46 <zhiyuan> well, i think our neutron unit test needs to be modified again
13:20:41 <joehuang> ok, this issue could be checked tomorrow
13:21:09 <joehuang> #topic cross pod L2 networking and dynamic pod binding
13:21:28 <joehuang> hello, after several rounds review
13:21:44 <joehuang> these two specifications are close to be merged
13:22:19 <joehuang> Yipei, what's feed back from Shinobu
13:22:54 <Yipei_> he gives some comments on the description, and i have already revised and update it
13:23:13 <joehuang> seems some more concrete comments and easy to address
13:23:17 <joehuang> no -2 now
13:24:25 <joehuang> ok, please keep update, if any issue, please talk to the reviewer and update in time, let's close the spec review soon
13:24:38 <Yipei_> ok, got it
13:24:58 <joehuang> and Ronghui, how about the cross pod L2 networking spec
13:25:33 <ronghui_1> i update the spec
13:25:42 <ronghui_1> and there is also -2
13:26:27 <joehuang> ok, please communicate with the reviewer
13:26:31 <ronghui_1> ok
13:27:04 <joehuang> and have you checked each comment from Shinobu, especially the comment lead to -2?
13:27:25 <ronghui_1> i checked
13:28:58 <joehuang> please talk to Shinobu to check to see if some improvement is still missing
13:29:05 <ronghui_1> i will mail to Shinobu and for the comment latest
13:29:11 <ronghui_1> ok
13:29:42 <joehuang> ok
13:30:23 <joehuang> we need to have the spec being approved before the end of this month, the sooner, the better
13:31:43 <joehuang> any comment on these two specs?
13:32:03 <ronghui_1> no
13:32:45 <joehuang> ok, how about the development of these two features
13:33:35 <zhiyuan> if xiongqiu and ronghui are not familiar with the network stuff, i can take the job to work on cross pod l2 networking
13:34:36 <ronghui_1> that good for us to work  and learn
13:34:50 <joehuang> I saw xiongqiu has submmited several patches, which are also very important to tricircle, good job
13:35:04 <longxiongqiu> i am not familiar with network, i can follow zhiyuan when he submit patch, and try to do some work
13:35:06 <Yipei_> still reading source code of nova scheduler, i still need more time to study it.
13:35:56 <zhiyuan> nova scheduler will interact with nova conductor, you can skip this part
13:36:16 <zhiyuan> you just see how the filter scheduler is loaded and used
13:36:44 <joehuang> Xiongqiu's way is a good way to dig into tricircle, the more you write code, the more you know how to work on it
13:36:51 <Yipei_> host_manager and filter_scheduler?
13:38:02 <zhiyuan> yes, yipei
13:38:20 <Yipei_> ok, i am studying it
13:40:02 <joehuang> ok, keep going
13:40:17 <joehuang> #topic policy
13:41:03 <joehuang> will submit a patch soon using policy for RBAC API check
13:41:36 <longxiongqiu> good
13:41:37 <joehuang> for API which will not directly forward to bottom OpenStack, policy based RBAC check is needed
13:42:20 <joehuang> if the api request will be forwarded to bottom openstack, the bottom openstack can check that
13:42:40 <joehuang> but this should be done case by case
13:43:22 <longxiongqiu> like tricircle/api ?
13:43:46 <joehuang> tricircle API should be checked based on policy
13:44:04 <joehuang> currently it's hard coded only admin role is allowed
13:44:14 <longxiongqiu> yes
13:45:25 <longxiongqiu> @joehuang, where can i learn RBAC API check code?
13:45:42 <longxiongqiu> like nova or cinder policy
13:46:08 <zhiyuan> policy is checked in the api controller
13:46:29 <joehuang> it's easy for there is one oslo.policy lib
13:46:41 <joehuang> don't worry about it
13:46:57 <joehuang> xiongqiu, you can check our todo list
13:47:25 <joehuang> and a lot of feature parity is needed in tricircle like vm/volume operations
13:47:57 <longxiongqiu> ok
13:48:19 <joehuang> #link http://docs.openstack.org/developer/oslo.policy/
13:48:33 <joehuang> ok, any other topics?
13:48:42 <longxiongqiu> no
13:48:44 <zhiyuan> no
13:48:46 <ronghui_1> no
13:48:48 <Yipei_> no
13:48:58 <joehuang> ok, good night
13:49:04 <joehuang> #endmeeting