05:01:35 #startmeeting servicevm-device-manager 05:01:36 Meeting started Tue Jul 29 05:01:35 2014 UTC and is due to finish in 60 minutes. The chair is yamahata_. Information about MeetBot at http://wiki.debian.org/MeetBot. 05:01:37 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 05:01:39 The meeting name has been set to 'servicevm_device_manager' 05:02:02 Today Bob doesn't attend 05:02:38 Do we have s3wong? 05:03:12 yamahata: I want to introduce vish from Brocade 05:03:19 seems no. anyway today we don't have much topics. 05:03:22 natarajk: sure. 05:03:28 he will be contributing to ServiceVM 05:03:31 #topic Announcement 05:03:50 #chair natarajk 05:03:51 Current chairs: natarajk yamahata_ 05:04:04 Hello everybody....I am Vish and work with Karthik at Brocade 05:04:17 vishwanathj: hi 05:05:02 yamahata: i have added review comments in the latest Tacker spec 05:05:25 natarajk: Sure, I'll check it today or tomorrow. 05:05:36 sorry, a bit late 05:05:57 I have some questions on it. I'd like to discuss it later 05:06:00 s3wong: Hi 05:06:03 i still see some requirements regarding physical topology support 05:06:04 no problem 05:06:43 natarajk: what kind of requirement? 05:06:51 in tacker spec 05:07:29 i couldn't completly understand the topology related requirements (It is documented under responsibilities section) 05:07:36 I see. I think attach/detach interface api is needed in additio 05:08:03 yes 05:08:31 #topic servicevm spec review 05:09:02 natarajk: which line of the spec? 05:09:27 line 54 05:09:43 is it physical topology of VMs ? 05:10:02 what is it used for ? 05:10:50 It's about service insertion/chaining 05:11:11 do we need to store it 05:11:13 ? 05:11:18 wait 05:11:56 Sorry, I misunderstood. It's about physical appliance case. 05:12:30 For physical appliance why do we want to discover the topology ? 05:12:39 Cisco CSR1k supports both physical/virtual. So they want uniform API for virtual/physical 05:13:34 would we use the topology information to do policy based allocation ? 05:13:50 natarajk: Correct. For that purpose. 05:14:19 i think we need to add some examples in the responsibilities section 05:14:32 natarajk: For routervm l3 plugin case, we have mostly none for it. 05:14:59 Bob proposed the sentence, and he'd like to clarify future direction. 05:15:00 yes, that's what i thought 05:15:10 we can continue the discussion in gerrit 05:15:16 natarajk: sure. 05:15:46 #topic openstack conference submition 05:16:06 natarajk: you have submitted the proposal. 05:16:12 yes 05:16:13 natarajk: thanks 05:16:36 welcome. Let's hope it gets selected :) 05:16:58 (Un?)Fortunately, the proposal is also submitted from Intel internal queue. 05:17:03 natarajk: yes, Thanks! 05:17:10 Later we will arbiterate 05:17:20 natarajk: Probably I'll contact you later. 05:17:31 yamahata: sure 05:18:15 #topic open discussion 05:18:54 yamahata: i saw that l3 db mixin refactoring is getting good comments 05:19:07 natarajk: Yes. Today I'll respin it. 05:19:19 #link https://review.openstack.org/#/c/108728/ l3 db mixin refactoring 05:19:39 natarajk: Do you have any requirement on refactoring? 05:20:03 natarajk: I hope it's also useful for vyatta l3 plugin 05:20:13 yamahata: in my l3 plugin, i am using the rpc notifier for the firewall agent 05:20:45 as firewall agent needs messages from l3 plugin 05:20:53 regarding router config updates 05:21:26 natarajk: I see. Do you have firewall driver publicly available? or not yet? 05:21:45 we have a POC 05:21:58 cool 05:22:39 but as firewall service plugin framework is getting refactored, i am waiting for it to get stabilized 05:22:49 natarajk: understand 05:23:00 I have one question on l3 plugin implementation 05:23:04 yes 05:23:15 RE: router-remove-interface 05:23:31 The neutron port is deleted on behalf of nova detach interface 05:24:15 Since the device owner of neutron port is router, so doen't port deletion fail? 05:24:41 Or you do some trick to work around it? 05:24:43 we change the owner during the port deletioin process 05:24:56 Oh. Got it. 05:24:59 you can check the code. 05:25:16 I missed it. I'll recheck the code. 05:25:27 we set the device_owner to empty before deleting the port 05:25:44 Makes sense 05:26:27 thanks for explanation 05:26:38 welcome 05:26:54 anything else to discuss? 05:27:49 okay thank you everyone 05:27:56 see you next week 05:28:13 bye 05:28:18 bye 05:28:26 #endmeeting