17:01:21 #startmeeting networking_l2gw 17:01:22 Meeting started Mon Jul 18 17:01:21 2016 UTC and is due to finish in 60 minutes. The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:26 The meeting name has been set to 'networking_l2gw' 17:01:45 can I see a show of hands who is here to attend L2GW meeting? 17:02:07 I'm here 17:02:35 I see bunch of guys on the channel - wonder who else is here 17:02:50 #topic: Agenda 17:02:54 #link: https://wiki.openstack.org/wiki/Meetings/L2Gateway#Meeting_July_18.2C_2016 17:03:12 #topic: Announcements 17:03:29 oferby is a new core member of this team 17:03:36 congratulations oferby 17:03:43 :-) thanks 17:04:08 #topic: Patches under review 17:04:21 #link: https://review.openstack.org/#/q/status:open+project:openstack/networking-l2gw,n,z 17:04:36 I approved some patches over the weekend 17:05:09 I noticed new ones have popped up this morning 17:05:24 I am glad to see oferby is here to help out with these 17:05:46 I add small change just for small cleanup 17:06:05 cleanup is always good :-) 17:06:19 if you are ok with this I will continue ... 17:07:16 oferby : I will review it later this evening and post comments 17:07:27 ok 17:07:39 also, I will be available on-line this evening, should you need to reach me 17:07:55 ok 17:08:14 #topic: OVSLIB 17:08:28 oferby : thanks for taking on this 17:08:53 one of my concerns is keeping this updated with the on-going changes 17:09:19 I totally understand and this was my concern too - 17:09:43 but could not convince OVS people ... 17:10:10 the issue is not that big - 17:10:29 the schema is used by IDL only to ask for tables and columns 17:10:32 I have just added russell as a reviewer to this patch - to seek his feedback 17:10:56 he was involved before Ben .... 17:11:11 is he not still involved? 17:11:34 stopped after Ben made his view on this 17:12:08 hmm.... I do not see him on IRC - 17:12:29 will try to ping him later during the day and ask him to provide feedback - 17:13:27 armax: are you here? 17:15:29 oferby : I will try to reach armax as well russell later during the day to see if we can get him to review the patch as well 17:15:44 ok. 17:15:54 I like what I saw - I think this is a good staring point to get things moving forward 17:16:20 I have few thinks that I would like to do in neutron-lib to help us - 17:16:37 it would be nice if you could ask armax to join. 17:16:56 I will 17:17:21 feel free to shoot him an email - if your timing on the IRC does not work out - 17:17:29 just for you to know what it is about - 17:17:49 there are some classes in neutron that we need - 17:17:58 Connection, Transaction, etc - 17:18:12 right 17:18:23 I don't want to copy them, I rather use small package like neutron-lib - 17:18:38 so I would like to move them from neutron to lib 17:19:00 I would like to have armax ok before doing the work ... 17:19:20 you view on it?! 17:19:21 why can't we simply import it? 17:19:44 because neutron is too big of a code and we need only fraction. 17:20:04 if you have a node with only the l2gw agent - 17:20:20 can't we just import the classes that we are interested in? 17:20:21 you will need to install neutron-server code. 17:20:32 no. dependencies ..... 17:20:41 ah I see 17:20:54 also installation script will not allow this. 17:21:14 I do not like the idea of running neutron server on the hypervisors 17:21:31 you do not run it 17:21:47 just install all the code to use small part of it. 17:21:55 silly .... 17:22:10 understood 17:22:49 but, I have a feeling, if you import neutron-lib anywhere, you are getting the code anyways 17:23:31 you don't get all neutron code which is big. only small part. 17:24:14 I mean the lib - not all of neutron server 17:26:53 let me follow up with armax and russell on this 17:27:01 will get back to you 17:28:08 #topic: L2GW as cloud inter-connect 17:29:12 oferby: looks like you were disconnected 17:29:20 yes I was... 17:29:40 so, on the cloud inter-connect 17:30:22 The spec looks good to me 17:30:36 I noticed few more reviews - which is always good 17:30:48 We should wrap this up - 17:31:08 right 17:32:53 did you get disconnected this time?! 17:33:08 I will give it a final read tonight and get this done 17:33:19 great. 17:33:20 probably 17:34:02 #topic: Open Discussion 17:34:35 oferby : anything else you would like to discuss? 17:34:48 I spoke with Swami earlier - 17:34:53 oh good 17:34:59 asked him to add comments on the DVR 17:35:02 Swami : are you here? 17:35:09 he promised he will today. 17:35:19 Sukhdev: yes here 17:35:32 Swami : we need to close off on that patch - 17:35:46 I just posted a couple of comments in there, once I get the answers I will be comfortable giving a +1 17:36:09 oferby : have you read the comments? 17:36:24 or lets quickly go over those comments here 17:36:35 ok. looking. 17:36:51 Swami : do you mind articulating your concern here - so that we can try to get a closure here 17:37:00 Sukhdev: sure 17:37:33 The question that I had was what does the endpoints contain? Does it have all the vxlan tunnel endpoints configured. 17:38:09 The next question is, if you are matching with the endpoint[host] to the agent['host'] how do you determine if this is the active host, if HA is enabled. 17:38:49 the endpoint contain host information. I look for the host name and resolve its IP address 17:39:19 oferby: ok, 17:40:15 oferby: so how do you determine, if this is the valid agent if HA is enabled for the network node, is there anyplace that you check for those conditions. 17:40:53 in HA case there will be multiple hosts and I'm selecting the first active one. 17:40:57 oferby: Or is that taken care somewhere else that is not in this patch. 17:41:20 oferby: so do you check for the active state of the agent in the network node. 17:41:39 yes. the db query return only active nodes. 17:42:28 oferby: you mean the filter with 'alive['True']' will take care of it. 17:42:36 right 17:42:43 oferby: ok, no problem. 17:42:56 I think I got what I wanted, I will +1 it right now. 17:43:02 oferby: Sukhdev: thanks 17:43:08 great. thanks. 17:43:24 oferby: jenkins is still failing. 17:43:30 Swami : you can post your comments based upon this discussion - 17:43:31 looking. 17:43:55 oferby : can you please take care of the py27 errors? 17:44:14 Once Swami gives a +1, and jenkins is happy, I will approve this for merge 17:44:20 I sure will. my tomorrow .... 17:44:29 sounds good 17:44:45 Swami : thanks for your time - much appriciated 17:44:54 Sukhdev: no problem 17:45:23 I see the problem. it is a fix for package that was missing in unit tests. 17:45:34 the fix was merged. 17:45:37 * Sukhdev owes Swami a beer in Barcelona 17:45:53 Sukhdev: sure, lets sync up 17:46:27 oferby : so, rebasing may take care of it then 17:46:39 it will. I will do that tomorrow. 17:46:53 your tonight... 17:47:01 sounds good - thanks oferby 17:47:18 anything else that we need to discuss? 17:47:22 nop. 17:47:29 good - so, we are done then 17:47:37 Thanks for attending 17:47:43 bye 17:47:50 ok. have a great day. bye. 17:48:01 #endmeeting