00:35:12 #startmeeting Yardstick Work Meeting 00:35:12 Meeting started Tue Mar 21 00:35:12 2017 UTC. The chair is rbbratta. Information about MeetBot at http://wiki.debian.org/MeetBot. 00:35:12 Useful Commands: #action #agreed #help #info #idea #link #topic. 00:35:12 The meeting name has been set to 'yardstick_work_meeting' 00:35:27 Kubi said he couldn't make the meeting 00:35:55 #topic roll call 00:36:16 #info Jing 00:36:42 #info Ross 00:36:50 Not sure how to start gotomeeting 00:37:29 I don't have the account to start GMT, either. 00:37:43 So may be IRC only. 00:37:54 okay 00:38:35 #topic action-item follows up 00:39:18 #info Jack 00:39:34 #info AP1 rbbratta will work with the MD5SUM and vivid 00:39:47 #info MD5SUM and vivid patches submitted 00:40:47 any other action items? 00:41:31 that's all, I think 00:41:42 #info CI status 00:42:14 #link https://build.opnfv.org/ci/view/yardstick/ 00:42:21 any update on CI? 00:43:01 Not so much so far, I will keep monitoring. 00:44:10 Okay 00:44:20 #info stable branch 00:45:15 what do we need for stable branch? maybe https://gerrit.opnfv.org/gerrit/31109 00:46:18 You mean cherry picking patch to the stable branch? 00:46:37 Yes, what do we need to cherry-pick to stable 00:46:58 supposed to be bugfixes only 00:47:37 Yes, all patches that fix CI 00:47:48 31109 looks like it is bugfix 00:48:12 Not sure about CPU/NUMA pinning support. 00:48:20 yes, this is a bugfix for apex scenarios. 00:49:14 which? CPU/NUMA is for Apex scenarios? 00:50:03 or 31109 00:50:11 I discussed with kubi, he thought CPU/NUMA should be picked to stable, but if anyone against, I will abandon them. 00:50:25 31109 is the bugfix 00:51:12 well I was working on a different method of calling ansible, but I don't have it ready yet 00:52:03 Using python? 00:52:35 Yes, dynamic inventory and extra-vars. It is code we use from another project internally. 00:53:12 rbbratta: Great! 00:53:33 I would also prefer we use ansible openstack modules with shade module, and run the openstack commands from the yardstick host. 00:54:59 But shade do not support some openstack options. 00:55:54 oh? do you have details? It depends what we need to do. 00:55:56 sorry, options->operations 00:56:51 JackChan: do you know what shade support now, a list maybe? 00:58:10 #link https://docs.openstack.org/infra/shade/genindex.html 00:58:11 But for stable branch, the question is risk, are the CPU/NUMA changes stable/tested, will they impact stable branch much? 00:58:14 Mingjiang: It is the documentation of all modules, We can check from this. 00:58:58 rbbratta they will not affect the stable branch. 00:59:04 we wouldn't use shade directly, just the ansible 2.2 openstack modules that call shade. 00:59:05 But the shade module do not support many operations. And some operations will be supported in ansible 2.3.0 00:59:35 Okay, so we can cherry-pick to stable branch 00:59:45 rbbratta: yes, some operations only supported in ansible 2.3.0 01:00:23 i see, we're only using the ansible openstack module 01:00:35 I am trying to find a example. 01:00:59 #link http://docs.ansible.com/ansible/os_nova_flavor_module.html 01:01:12 see extra_specs option 01:01:42 #info CPU/NUMA pinning okay to cherry-pick to stable 01:02:31 anything else for stable branch? 01:02:41 also it may require some extra libs to be installed on the node/host if we use openstack module, is it right? 01:03:03 Mingjiang: ansible should run from the docker container 01:03:40 I think MD5 and vivid patches are ready to merge today and will be picked to stable branch 01:03:48 Mingjiang: Maybe we should install shade on the host. 01:04:19 Yes, we should install shade in the docker container 01:04:58 rbbratta: i mean the host/node to be configured may also need something to be installed 01:05:23 #info MD5 and vivid patches are ready to merge today and will be picked to stable branch 01:05:48 Mingjiang: shade should use openstack rest API, there should be no need to touch the nodes 01:06:47 I lost track, what topic are we on? MS8, MS9? 01:07:31 rbbratta: Thanks for your info, if it only use openstack API, i think it's a clean and clear way 01:08:21 rbbratta: got it. 01:08:24 rbbratta: no new topic typed in 01:08:37 #topic MS8, MS9 01:08:51 #info ACTION: please update the the fix version of JIRA tickets which you won't implement within Danube 1.0 01:09:15 #undo 01:09:15 Removing item from minutes: 01:09:24 #action please update the the fix version of JIRA tickets which you won't implement within Danube 1.0 01:09:57 anything else on MS8, MS9? 01:11:07 #link https://jira.opnfv.org/issues/?jql=project%20%3D%20YARDSTICK%20and%20fixVersion%20%3D%20%22Danube%201.0%22%20and%20status%20not%20in%20(Closed%2CCancelled%2CResolved) 01:11:40 i've seen 49 01:11:48 49->47 tickets 01:12:16 i'll update it today asap 01:12:38 I will also update it today. 01:12:50 Me too 01:14:36 okay, next topic 01:14:50 #topic others 01:15:16 Merged yardstick: Bugfix: netperf_bottleneck execute failed because is_same_heat_context return True https://gerrit.opnfv.org/gerrit/31025 01:15:21 I'm working on add Centos support for docker, and VNFs 01:15:46 Deepak should have some more patches to submit soon, Ixia support and other stuff. 01:16:17 For Centos support I am converting everything to ansible, should have a draft soon. 01:17:22 look forward to the centos support 01:17:47 Also Deepak should be adding apachebench/nginx support for traffic generator 01:18:31 Then also eventually adding KPI support for barometer and ceilometer 01:18:44 So we need to make some Jira tickets. 01:19:41 great 01:20:27 #info Ross working on adding CentOS support for docker and VMs 01:21:38 I'm having problems with shade authenticating inside the docker though, something with identity_api_version. Hard to debug. 01:22:43 okay, end meeting? 01:22:58 okay 01:23:03 #endmeeting