17:04:29 #startmeeting networking_l2gw 17:04:30 Meeting started Mon Jun 8 17:04:29 2015 UTC and is due to finish in 60 minutes. The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:04:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:04:34 The meeting name has been set to 'networking_l2gw' 17:04:54 #topic: Agenda 17:04:59 #link: https://wiki.openstack.org/wiki/Meetings/L2Gateway 17:05:19 sorry folks jumping from one meeting to another 17:05:43 armax marutikamat : did you have a chance to look at the agenda? 17:05:55 yes 17:06:09 Yes I did 17:06:16 Are we good with it - or want to cover anything else? 17:06:31 #topic: Announcements 17:06:33 I will add something in the end 17:06:52 marutikamat: cool 17:07:05 The presentation at summit went very well - I got lots of queries - 17:07:14 I am sure you got many as well 17:07:29 Yes 17:07:29 #link: https://www.openstack.org/summit/vancouver-2015/summit-videos/presentation/bringing-provider-networks-into-openstack-using-l2-gateway 17:07:51 I was hoping that we will have many additional people joining us today 17:08:13 anybody in this meeting who has not introduced themselves yet? Please do so…now... 17:08:57 #topic: L2 GW Wiki updated 17:09:10 I took some time to update the wiki yesterday 17:09:45 #link: https://wiki.openstack.org/wiki/Neutron/L2-GW 17:10:14 Please have a look and feel free to update as needed - I also added links to the demos 17:10:33 #topic: Requested features in Liberty 17:10:40 Thanks Sukhdev 17:10:44 thanks Sukhdev, it looks good 17:10:56 I have listed one new feture 17:11:13 OVSDB server connection initiation - 17:11:32 I briefly chatted with you about it in Vancouver 17:11:49 Sukhdev: I wonder if we should have a ‘roadmap’ section that shows what we do/did each release 17:11:53 We need ability to be able to initiate connection from either side 17:12:09 armax: excellent idea 17:12:26 +1 with armax 17:12:41 armax: you want to add, or you want me to do the honors? 17:13:06 marutikamat: that goes for you too - I did not mean to exclude you 17:13:09 I got the idea Sukhdev what you are asking. However, we will need a BP for this. 17:13:31 Sukhdev: I can sketch something out? 17:13:48 armax: cool thanks - let me know if you need any help from me 17:13:49 marutikamat: either a bp or a bug fix 17:13:57 marutikamat: I mean bug report 17:14:29 I would think a bug should be sufficient - unless you want to go through the hassle of BP 17:14:31 Per the original spec, all the connections are opened by the gw agents. But in this change, the connections will be opened by the OVSDB server itself. 17:14:40 I could open a bug, if you like 17:15:30 ok 17:15:52 marutikamat: OK - as in OK to open a bug? :-) 17:16:05 Are we giving an option to the use in ini file? 17:16:20 s/use/user 17:16:31 marutikamat: yes, it will better - but, I am OK either way 17:16:38 please file a bug 17:16:49 marutikamat: will do 17:17:07 ok, we will figure out after investigation 17:17:25 armax marutikamat : any other feature you are considering for Liberty? 17:17:51 We’ve been saying we wanted a softweare l2gw 17:18:00 but no-one is stepping up and volunteer 17:18:11 they all want the cake but not pay for it :) 17:18:20 armax: correct :-) 17:18:33 Yes. Just to inform you, myself and selva have been pulled out and put into some other projects. I was also thinking of support for service node and supporting l2gw through ODL 17:18:49 armax: I am a bit disappointed - I was hoping to see lot more new people here asking for all kinds of features 17:19:09 armax i will send out a separate e-mail on the l2gw-ODL integration 17:19:15 marutikamat: ok, let’s sync up offline 17:19:40 Sukhdev: perhaps we need more advertising, during the neutron irc meeting? 17:20:07 armax: I did few times in the past - I'll put it on next week's meeting 17:20:27 nsx supports service node for unknown traffic. we should also support it. 17:21:25 marutikamat: you mean punt all BUM traffic to service node? 17:21:49 one more item s DVR with l2gw 17:22:02 we should start building a backlog of items, via bug list and whatnot 17:22:04 yes Sukhdev 17:22:49 armax: Actually, we can list them on the wiki - when you list roadmap, perhaps we can add a section for future (long term) items - 17:23:02 armax: that may generate interest among others 17:23:12 Sukhdev: I was thinking that the roadmap should capture big ticket items 17:23:23 and duplicating the list is a bit of a pain to keep in sync 17:23:57 armax: we can create a bug and put the pointer under roadmap - that will keep things easy 17:24:21 marutikamat: can you open bugs for each of the items you mentioned? 17:24:31 ok sure Sukhdev. 17:24:42 please put action item on me 17:24:43 Sukhdev: ok, let me see what we put together 17:25:20 Is Isaku there in the meeting? 17:25:21 #action: marutikamat to create bugs for each of the new feature enhancements for L2 Gateway 17:25:38 marutikamat: I do not see him here 17:25:52 He was keen in software l2gw. 17:26:24 yes, he expressed interest, but, never came forward with anything concrete 17:27:22 #action: Sukhdev to open a bug for OVSDB connection initiation from OVSDB server 17:27:32 #topic: Bugs 17:27:58 marutikamat: I had discussed with you about the flood list issue 17:27:59 Sukhdev, people are facing problems while backporting the fixes 17:28:16 marutikamat: what is the issue? 17:28:37 armax: is expert in this kinds of issues - I usually bug him when I get stuck :-):-) 17:28:38 Our test enginners (Ashish and Alok) are able to ping from bare metal server to the VM first. 17:28:40 but, I can help 17:29:19 vikas sent you an email on that. the branch shows as master even if he is trying to commit on kilo/stable. 17:29:23 marutikamat: after the first ping from VM or adding arp entry solves it 17:29:57 they do not ping from VM first. The first ping itself from BM to VM works for them. 17:30:03 marutikamat: are you saying that this is a regression? 17:30:17 marutikamat: or in other works did the test case ever pass? 17:30:21 I was expecting their email reply to you but they missed it out today, 17:30:26 marutikamat: the best way to do is to go on the review board - open the patch which merged in master - and click on a tab called cherry pick - click on it and rest is easy 17:30:27 I will check with them again tomorrow. 17:30:48 if so, a good tool to use would be git bisect 17:31:36 thanks to both of you for the suggestion 17:31:48 marutikamat: I think the best way to go about is to dump the OVSDB contents and paste it - I sent you the paste from my side 17:31:59 marutikamat: that’s a handy and quick tutorial http://webchick.net/node/99 17:32:06 Sukhdev: I will request Alok/Ashish to get in touch with you for the ping problem 17:32:10 marutikamat: I see that HS VTEP publishes the flood list and SW VTEP does not 17:32:21 marutikamat: basically ‘bisect’ will help you identify what commit broke the feature 17:32:23 marutikamat: cool - 17:32:39 and then it’s easier to understand what’s going, that assumed that it is a regression 17:32:46 armax: thanks 17:33:47 armax: thanks for sharing this tool - I did not know about it 17:34:07 Sukhdev: sure thing 17:34:23 As to the bugs - I approved few 17:34:42 I will review more and approve them this week 17:34:52 thanks Sukhdev 17:35:05 we are looking good from the bugs point of view 17:35:22 marutikamat: I have exhausted my agenda items 17:35:32 you mentioned you wanted to bring something ? 17:35:41 #topic OpenDiscussion 17:35:51 Lets hear it 17:36:33 I already mentioned above. Additional backlog items 17:36:48 Oh I see - cool 17:36:48 I will submit bugs for those 17:36:55 good 17:36:59 anything else? 17:37:13 nothing from me 17:37:23 Looks like we are done - 17:37:32 thanks folks 17:37:46 thanks both of you 17:38:02 #endmeeting