08:05:29 #startmeeting multisite 08:05:29 Meeting started Thu Dec 3 08:05:29 2015 UTC. The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:05:29 Useful Commands: #action #agreed #help #info #idea #link #topic. 08:05:29 The meeting name has been set to 'multisite' 08:05:45 #info dimitri 08:05:54 #info joehuang 08:05:57 #info Ashish 08:06:02 #info Malla 08:06:19 #topic documentation 08:06:45 do you know whether there is user guide and configuration guide template 08:07:08 I would like to write the documentation for some use cases from multisite 08:07:32 Officially, there is no template 08:07:52 So I guess we can write in a simple free form document and send it to Chris 08:08:00 He’ll do all the formatting 08:08:05 Other projects are using their own template as I know 08:08:12 hi, 08:08:14 If so it would be simple to do 08:08:23 hi, midu 08:08:23 we now have template 08:08:29 sorry, mibu 08:08:45 #link https://wiki.opnfv.org/documentation/tools 08:09:09 just fyi :) 08:09:28 thanks mibu.. :-) 08:09:29 great, I searched but not found 08:09:52 thanks a lot mibu :) 08:10:16 we still have to create policy of documents types, so I'm asking chris to fix it 08:10:42 after that he or I will send emailto announce it 08:10:44 if you would like to write documentation, please feel free to pick one jira ticket 08:11:07 when will it be planned to be ready? 08:11:29 not sure... 08:11:40 but you can create your docs first 08:11:48 good idea 08:12:08 it only two months left for B release 08:12:14 it is easy to move source dontents to rearrage doc structure 08:12:47 yes, I faced the same problem in doctor, so I created/updated toolchain 08:13:04 great helpful 08:13:11 let's work together :) 08:13:17 sure 08:13:25 #topic kingbird 08:14:03 to Ashish 08:14:09 yes. 08:14:33 Zhiyuan said that openstack client does not support all features especially for neutron 08:14:53 Do you have any idea about it 08:14:58 yes.. It is still evolving 08:15:20 and I have seen it growning from recent releases. 08:15:30 growing* 08:15:42 will it affect the feature we want to implement 08:15:49 hi zhiyuan 08:16:11 could you describe why you use keystone client instead 08:16:17 in the endpointcache 08:16:30 hi, I went to openstack meeting and found no one 08:16:38 yes. It may as of now, But as a by product of our activity we may also contribute to OSclient 08:16:53 sorry it's in OPNFV channels 08:18:25 so we improve OSclient? it lacks support for most of the resources in Neutron, 08:19:03 somewhere I read keystone client will be deprecated and openstackclient will be used instead 08:19:17 same I think will be with other native clients 08:19:17 Is there anyone actively contributing in the Neutron client 08:21:01 I am not sure on that, But have seen the releases happening for OSclient very frequently and its improving. 08:21:04 if there is no one actively contributing in openstack client, and also neutron client, then we may have to use neutron client for a while. It's up to whether it will affect our feature 08:21:06 as I know, the http client part of keystone-client is still in use, but shell part is moved to openstack-client. 08:21:39 ok let's move to openstack client first 08:21:49 so it's recommended to use openstack client if in terminal 08:22:06 only if some feature is not ready in OpenStack client then will use each service client temaryly 08:22:46 yes.. I agree its very much cohesive to CLI, But we can write python bindings for it. 08:22:52 we should first clarify if using openstackclient the way want to use it is right 08:23:20 to me they way we create its instance doesn’t look very nice. we’re basically emulating CLI 08:23:56 do you want to issue rest api request directly? 08:23:59 yes.. the argparse and all.. 08:24:27 if so we can try the python SDK 08:24:53 No, Then there is already native APIs 08:25:01 novaclient, keystoneclient 08:25:22 yes, so the question is will those clients be deprecated 08:25:36 keystone-client will be 08:25:54 event the http binding? 08:26:48 I think openstack-client is just a wrapper, for endpoint cache, use keystone-client is fine, since openstack-client is also based on keystone-client. For resource operation, I once considered to use openstack-client but it lacks support on Neutron so I gave up 08:26:51 hmm.. Need to check. 08:27:41 so in another project I work on, Tricircle, I write a wrapper myself 08:27:47 not having support for neutron will cause a lot of problems 08:28:18 there were problems with quota part too as far as i remember 08:28:19 as Joe said, we will use native clients as of now 08:28:28 yes.. updating few quota resources 08:28:32 it shows problem 08:29:21 native clients as of now for the features which are not present in openstackclient 08:29:38 ok 08:29:42 for all supported features, we will use OSclient 08:29:49 yes we have to do so 08:29:50 My suggestion 08:30:14 #agree native clients as of now for the features which are not present in openstackclient 08:30:23 #agree for all supported features, we will use OSclient 08:31:24 to Ashish, how about the testing execution if using tox -epy27 08:31:52 the test has been done successfully in CI gate check 08:31:54 there’s the run_tests.sh script 08:32:04 Yes. Joe, Still some problems, I think they code for py34 it is trying to execute with py27. I am getting syntax errors 08:32:22 it runs all units tests and also styling checks 08:32:31 Compiling /root/soni_verify/kingbird/.tox/pep8/build/pylint/pylint/test/functional/abstract_abc_methods.py ... File "/root/soni_verify/kingbird/.tox/pep8/build/pylint/pylint/test/functional/abstract_abc_methods.py", line 6 class Parent(object, metaclass=abc.ABCMeta): ^ SyntaxError: invalid syntax Compiling /root/soni_verify/kingbird/.tox/pep8/build/pylint/pylint/test/functional/abstract_ 08:32:33 I did not try run_tests.sh 08:33:13 yes.. run_test.sh works fine 08:34:11 with tox I am getting these issues. 08:34:28 gate-kingbird-python27 SUCCESS in 1m 53s https://review.openstack.org/#/c/250707/ 08:34:41 let me see 08:35:08 sure. 08:35:41 you mean the test cases execueted successfully if run_test.sh, but failed "tox -epy27" ? 08:36:23 yes.. 08:36:43 kingbird.tests.db.test_quota_db_api.DBAPIQuotaTest.test_update_quota_limit 0.063 kingbird.tests.db.test_quota_db_api.DBAPIQuotaTest.test_create_quota_limit 0.034 kingbird.tests.db.test_quota_db_api.DBAPIQuotaTest.test_delete_quota_limit 0.031 kingbird.tests.db.test_quota_db_api.DBAPIQuotaTest.test_delete_all_quota_limit 0.030 kingbird.tests.functional.api.testroot.TestKeystoneAuth.test_auth_enforced 0.012 k 08:37:07 sorry, It's not visible properly. 08:37:15 got it 08:37:17 try\ running ./run_test.sh 08:37:25 I'll check tomorry 08:38:04 sure.. 08:39:08 any other information to share? 08:39:16 or issue to discuss? 08:39:34 The specs for BPs are updated 08:39:38 have you checked those? 08:39:48 for CQM and Generic driver 08:40:03 So, the BPs. How about moving the google docs to kinbirdi openstack wiki? 08:40:06 I'll review it next few days 08:40:14 no problem 08:40:28 to sorantis: no problem 08:40:33 sure. Guys if anybody is facing issues with OSclient. Let me know, I will try to help 08:41:05 to Ashish: thanks update. review later 08:41:16 SAshish, we agreed that osclient need to be changed to service clients 08:41:22 sure 08:41:44 so will you make this change in the existing commit or abandon it? 08:41:53 to soranti: if osclient not support features we require 08:42:19 yes.. See right now authentication with each services is happening properly 08:43:04 If in future we use any feature which is not not supported then we will introduce native client, So I think the current commit need not change 08:43:34 if there are any other comments, then yes can do that. 08:44:00 #action check tox error 08:44:01 Am I making sense? 08:44:06 yes 08:44:10 thanks :) 08:44:21 my only concern is that we do know that osclient has problems with quota, no neutron support and currently endpointcache is using keystone client 08:44:29 so why do we keep it? 08:45:46 as it looks so far the only change will be in the sdk.py 08:45:47 yes. I agree but It will improve and if not now, we have to use it in future. 08:46:24 ok 08:46:45 is everybody ok with that? 08:46:52 ok 08:46:55 ok 08:47:53 I don't think it's good idea some one of us to fix the bug of osclient or add new feature in it. we need to focus on kingbird itself 08:48:05 +1 08:48:15 we have very limited resources and time 08:48:41 we can always submit a bug though 08:48:45 or a feature request 08:48:49 +1 08:48:53 agree 08:49:09 basically that’s how collaboration with other projects begins 08:49:18 +1 08:49:41 +1 08:49:41 and we work for a collaborated project 08:50:04 project => platform 08:50:18 #info we can always submit a bug though or a feature request, but we need to focus on kingbird itself 08:51:24 let's work offline and conclude today's meeting if no more topic 08:51:31 yes 08:51:34 yes 08:51:39 thanks for your attending 08:51:43 thanks! 08:51:46 #endmeeting