07:06:01 #startmeeting multisite 07:06:01 Meeting started Thu Sep 8 07:06:01 2016 UTC. The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot. 07:06:01 Useful Commands: #action #agreed #help #info #idea #link #topic. 07:06:01 The meeting name has been set to 'multisite' 07:06:11 #topic rollcall 07:06:17 #info joehuang 07:06:23 #info Ashish 07:06:27 #dimitri 07:06:32 #info dimitri 07:06:47 #topic next step open discussion 07:07:15 are you people able to access the lab? 07:07:21 yes 07:07:49 after they changed the firewall rule, don't know which rule 07:07:49 when did you try lately? 07:08:16 last week or the week before that 07:08:20 what's the issue 07:08:30 currently its not accessible 07:08:34 https://build.opnfv.org/ci/view/multisite/job/multisite-kingbird-daily-master/ 07:08:59 but I can access it locally 07:09:05 access the node 07:09:08 locally 07:09:35 daily job triggered recently on 2nd 07:09:48 since then there is no run 07:11:17 so the openstack are not working> 07:12:02 can you access the node now 07:12:30 I am not able to login to setup 07:13:03 sorry. I am able to 07:13:04 no VMs in the node 07:13:34 [jenkins@pod4-node6 ~]$ ssh root@10.20.0.2 07:13:37 ssh: connect to host 10.20.0.2 port 22: No route to host 07:13:38 all fuel and MOS were erased away 07:13:51 is there any update? 07:14:17 I have logged on the the node, but nothing on it 07:14:36 [root@pod4-node6 ~]# virsh list 07:14:36 Id Name State 07:14:37 ---------------------------------------------------- 07:15:12 I’ll ask what happened 07:15:15 even no docker installed 07:15:28 [root@pod4-node6 ~]# docker ps 07:15:29 Cannot connect to the Docker daemon. Is the docker daemon running on this host? 07:16:42 something must have happened to the lab. or they didn’t inform us about lab work 07:16:57 yeah, may be some update or so 07:17:27 ok, will consult in the mail-list 07:17:28 this practically means that we can never assume our node stable 07:18:25 so your suggestion? 07:18:54 unless we automate the m-r installation there’s nothing we can do actually 07:18:59 OR 07:19:18 if anything happens or about to happen with the node, we should be informed 07:19:34 they didn't 07:19:40 even we have 1, we should have 2 as well 07:20:03 we may have to run tempest in single region in OpenStack 07:20:29 so, do we have any other node 07:20:29 ? 07:20:36 no 07:20:39 to make sure test cases runing for new patch 07:20:57 but they must inform the user 07:20:59 :( 07:21:53 this is very bad 07:23:32 please review the patch: https://gerrit.opnfv.org/gerrit/#/c/20635/, OPNFV hope all license update to creative common 4.0, and we need to update immeditly 07:24:17 emailed 07:24:59 thanks 07:25:43 last meeting, we discussed some topic to work in release D 07:25:56 1)KB CLI (joehuang, 07:26:12) 07:25:56 2)KB SDK (joehuang, 07:26:21) 07:25:56 3)Resources Sync (joehuang, 07:26:37) 07:25:56 3.1)SEG (joehuang, 07:27:36) 07:25:56 3.2)FLAVOR (joehuang, 07:27:54) 07:25:58 3.3)SSH KEY (joehuang, 07:28:08) 07:26:01 Revisit Geographical Redundancy 07:26:22 4)cross neutron networking autom 07:26:33 5) installation cooperation (joehuang, 07:32:18) 07:27:11 ther are some topics which are more clear, just to implement 07:27:27 some topics needs discussion 07:28:23 let's schedule these topics for which needs discussion 07:29:45 your proposal? 07:30:50 I’ve already started som work on the kingbird client 07:31:03 I am looking into SEG syncing 07:31:04 yes, saw that, it's much more clear 07:31:18 since it’s pretty straightforward, I think there’s not much to talk about 07:31:31 implementation is standard to other openstack clients 07:31:43 we just need to ensure that the APIs are properly invoked 07:31:44 we shall complete kb CLI, then can move forward 07:31:48 yes, for client, it's quite straightforward 07:32:36 #info KB CLI is pretty straightforward, just implement it. 07:33:10 for SEG, discussion is needed 07:33:10 regarding kb SDK. I think the priority to this is quite small, so we can consider the other points before coming to this one 07:33:53 can kbSDK included in KB CLI 07:34:09 ususally CLI also call the SDK 07:34:19 I think so. I haven’t looked into it yet. but I guess it’s just another wrapper to the APIs 07:34:29 I’ll look into it 07:34:42 correct, just a client wrapper, like novaclient class 07:34:43 we include kb cli to openstack sdk 07:35:26 ok, time to merge the initial patch 07:35:31 it seems that the requirements are out of date 07:36:06 not include kb cli to openstack sdk, I think it's not possible if KB is not an offical project to be included in openstack sdk 07:36:29 we shall see 07:36:41 we can always write a plugin that connects to openstackcli 07:37:05 #action check whether it's possible to include kb cli/sdk in openstack 07:37:24 one doubt guys 07:37:30 please 07:37:49 how kbsdk is diff from kb cli? 07:38:12 I thought kbsdk == inclusion of kb cli in openstack sdk 07:39:40 kb cli with which we get cli and kb python binding 07:39:43 kbsdk is to provide a client package which could be called by other package 07:40:33 when you install openstack client, cli provided 07:40:38 at the same time 07:41:00 when we install nova client we get python bindings which can be called by other packges also right 07:41:09 + we get CLI 07:41:21 other package for example kingbird can import client and write code to do the api calling 07:41:24 yes 07:41:28 but now it’s separated 07:41:40 since the intention is to have a uniform client for all services 07:41:43 oh I see 07:41:50 that is openstack sdk 07:41:52 not kb sdk 07:41:54 right? 07:42:29 openstacksdk is a way to have programmatic access to all services with a single authentication 07:44:22 provide kb sdk is important, programmatic calling will be more practical than manaullay using cli 07:46:11 hello, dimitri, not received your mail in the mail-list? 07:46:25 I email it to my coleague 07:46:29 He’s here now 07:46:36 I’ll ask him directly and let you know 07:46:45 btw, just updated the common layer patch, please review 07:46:46 good! 07:46:52 ok 07:47:13 so let's finish the kb client first, then discuss other topics 07:47:54 ok 07:48:01 so power went down in the intel lab over the weekend 07:48:28 Fatih will reinstall the node, and then I’ll reconfigure to m-r 07:48:39 thanks a lot! 07:49:20 other topics? 07:49:51 i guess for the time being we’re set 07:50:33 SEG 07:51:43 for what? SEG 07:51:53 Ashish wanted to talk about it 07:51:59 please 07:54:01 ashish? 07:54:07 hello, ashish? 07:54:24 yeah, as we discussed we will leave the remote security groups 07:55:39 and can sync others. 07:56:03 so I am still working on it, 07:56:21 are you working on a solution proposal? 07:56:31 for SEG, shall we implement it in KB or Tricircle(which only deal with networking part after spliting) 07:56:34 yes, but need some setup to verify 07:56:49 yes 07:57:18 and I also discussed with Bin, about how Gluon and Tricircle work together for some issues discussed in the mail-list which are not touched in gluon yet 07:57:24 joe, let’s first see what limitation we will face when syncing security groups 07:57:24 yes, will send the plan for SEG. but verification needs to be done 07:58:07 this should be part of the study of course 07:58:21 ok, more detail proposal, except that, in multi-region deployment, IP address space management and ip allocation is also needed to be considered 07:58:35 good! 07:58:44 we can have open discussion 07:58:55 yeah, need the setup soon 07:59:45 great 07:59:49 time’s up gents 07:59:54 yes 08:00:02 thank you for the meeting 08:00:04 Please follow up with Fatih 08:00:06 for setup 08:00:27 great appreciation 08:00:43 yes 08:00:46 yes 08:00:52 #endmeeting