16:01:34 #startmeeting networking_ml2 16:01:35 Meeting started Wed May 7 16:01:34 2014 UTC and is due to finish in 60 minutes. The chair is rkukura. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:38 The meeting name has been set to 'networking_ml2' 16:02:17 #link https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_May_7.2C_2014 16:03:12 #topic Announcements 16:03:38 Juno Summit next week, so no IRC meeting 16:04:20 Also, if you don’t attend the Neutron IRC meetings, you should be aware a mid-cycle sprint meeting is being planned for Juno 16:04:48 options being considered for date and location are at https://etherpad.openstack.org/p/neutron-juno-mid-cycle-meeting 16:05:50 I think refactoring the neutron core is part of the agenda for this, and ML2 involvement in that would be important 16:06:17 Any questions/comments on these announcements, or any other announcements? 16:06:48 Will ML2 also be affected by the task-based workflow plans? 16:06:49 if not, we’ll move on… 16:07:12 HenryG: good question 16:07:33 I haven’t looked at any specifics of that yet. What do you think? 16:07:54 HenryG: elaborate, please 16:08:08 Don't know yet. I plan to attend the task-flow session and see. 16:08:50 HenryG: when and where is the session? 16:09:15 Depending on the granularity of tasks being addressed, it could potentially be useful for postcommit processing, and maybe for error recovery 16:09:30 Sukhdev: I think is Tuesday lead by markmcclain1 but there is not BP yet 16:09:45 so it is a mystery right now 16:09:53 I have not seen a BP for the session yet. The session proposal is http://summit.openstack.org/cfp/details/433 16:10:29 I have been trying to dig it up - 16:10:50 defenitely looks relevent to ML2, and could result in significant changes to driver APIs, etc. 16:11:12 Anything else related to announcements? 16:11:17 Info on taskflow here: #link https://wiki.openstack.org/wiki/TaskFlow 16:12:02 thanks HenryG 16:12:07 #topic Action Items 16:12:38 the only official AI was for me, regarding etherpads for our summit sessions 16:13:03 I added one for the ML2 roadmap itself, and several have been added by others 16:13:48 We’ve got 3 sessions that are mainly focused on ML2, each incorporation several session proposals 16:14:17 One question is whether we should try to do a single etherpad per session, vs. etherpads per session proposal? 16:14:22 Any thoughts? 16:15:00 rkukura: single one makes easier to follow up action items 16:15:08 +1 for etherpads per session proposal 16:15:33 rkukura: earlier we thought about going with one per session - either way is OK 16:15:48 even though the subsessions are related, it may be difficult to bring them into one place. May be multiple leading to sthe session and unifying we get there? 16:16:27 We could certainly start with one etherpad per session, but with a section for each proposal 16:17:51 rkukura: that would be fine imo 16:17:58 Sounds like a good compromise. At least to get started. 16:18:05 rkukura: I think that is good 16:18:07 I think etherpad per session would help foster dicsussion across the entire topic rather than each propsoal separately 16:18:25 makes sense 16:18:25 rkukura: how much details regarding a desing should be added in the etherpad. 16:18:48 Would someone like to volunteer to pull together the etherpads for each of the three sessions? 16:19:20 I can do it for the modular agent session 16:19:22 nlahouti_: I think a page worth of bullet points per proposal is about the max 16:19:30 banix: Thanks! 16:19:43 I’m happy to do the ML2 roadmap session 16:19:53 rkukura: we have only two etherpad as of now 16:20:05 do we have the third one created? 16:21:02 someone put together https://etherpad.openstack.org/p/hierarchical_network_topology 16:21:45 This has sections for each of the proposals that went into it 16:21:55 rkukura: thanks for pointing this one - I was missing this…this looks good 16:22:12 but we probably should flesh-out each with some bullets to explain the idea, raise issues, etc 16:22:20 rukkura: Yes, this has the three sessions for Friday 16:22:54 We need to beef this one up 16:22:59 we’ve moved into the next topic... 16:23:12 #topic ML2 Design Summit Sessions 16:24:25 The agenda page has links to the neutron schedule, the 3 scheduled sessions, and the wiki with the etherpads 16:25:06 I’d like to see the session descriptions updated to be more clear, etc., and can ask mestery about that 16:25:49 Are there any ML2 session proposals were supposedly accepted/merged, but don’t have a home on one of these three session? 16:26:07 rkukura: I pointed it to him about one session - he mentioned you need to do it or coordinate with him :-):-) 16:26:31 Sukhdev: OK, if I can, I’ll try 16:27:12 rkukura: there was a session approved for inclusion but i dont see it in the list. the session on having a unified mechanism driver for SDN controllers. I do NOT think we need to have this but wanted to get clarification on it. 16:27:40 banix: I thiink the intention was to cover that as part of the ML2 roadmap session 16:28:08 rkukura: yes that was my understanding. 16:28:39 rkukura: ML2 Roadmap description is not very clear - this is what I asked Kyle to fix 16:28:39 banix: Lets put the link to the BP/spec on the etherpad for that session, and try to identify any dependencies, and who is working on it 16:29:10 Sukhdev: OK, I’ll see if I can fix that one, and if not, will work with mestery on it 16:29:28 rkukura: i think this is a simple and limitted of a topic that may not need a session; will follow up later. 16:30:07 banix: That was the goal for the roadmap session proposal - just to raise awareness of efforts that didn’t need their own sessions 16:30:28 rkukura: ok; makes sense. 16:31:10 This is the one for ML2 roadmap - need to beef up https://etherpad.openstack.org/p/ML2_Juno_Roadmap 16:31:24 There were two sessions that we approved that mestery marked as “merging with Isaku's session” 16:31:39 Do we know which session that is? 16:32:03 The proposals are http://summit.openstack.org/cfp/details/395 and http://summit.openstack.org/cfp/details/409 16:32:11 yeah hold on 16:32:27 These seem agent-related, but I’m not sure if they are in the ML2 agent session, or a different session 16:33:26 yes the above sessions were the ones 16:33:46 not in modular agent session 16:34:18 banix: So are these in a separate session? 16:35:06 this one? http://junodesignsummit.sched.org/event/9a0726f0946e4682bc77427afab44116#.U2pgnF7nwoc 16:35:29 not sure. 16:35:36 #action banix to organize single etherpad for Modular L2 Agents session 16:36:08 yamamoto: Yes, that looks like it. Thanks! 16:36:20 yeah yamamoto is right; i think that is the session^^^ 16:36:45 #action rkukura to flesh out single etherpad for ML2 Roadmap session 16:37:14 And the hierarchical session already has a single etherpad 16:37:48 per-proposal etherpads will be removed? 16:37:51 So each session proposer should try to flesh out an outline for discussion at the summit either the appropriate etherpad 16:38:49 yamamoto: If the assignees could move the per-proposal content to the the single etherpad, I think that the links to those can be removed - any reason not to? 16:39:39 no reason i can think of. 16:39:57 please try to think about how to get the most value out of these summit sessions 16:40:38 yeah i think for each session the stake holder need to talk and organize as how to present and direct the discussion 16:40:53 s/holder/holders 16:41:44 so this would be the main link : https://etherpad.openstack.org/p/ML2_Juno_Roadmap for the ML2 road map. 16:43:00 nlahouti_: that is presently listed under neutron etherpads 16:43:29 rkukura: some session, e.g. hierarchical networks has three topics...so each topic owner gets around 13 or 14 mins? If every topic owner in a etherpad session brainstorms before, it can be presented by one single person after collecting the ideas 16:43:31 nlahouti_: I will add content to this later today or tomorrow 16:43:49 sukhdev: yes it is under ' ML2 Juno Roadmap' 16:44:14 sukhdev: ok 16:44:22 padkrish: I think it is a good idea - 16:44:58 padkrish: I’m happy for the proposers in each session to figure out how they want to run that session 16:45:08 padkrish: I was thinking we could get together on Monday or Tuesday in Atlanta and hash it out, thoughts? 16:45:45 sukhdev# that's also good...i have proposed a webex session for tomorrow specifically for the hierarchical network session 16:45:54 Remember that these are supposed to be more discussion than presentation, so I don’t think having a single presenter is very imporant 16:46:09 padkrish: can you share the webex info? 16:46:25 sukhdev# sure... 16:46:41 padkrish: I was going to look at my calendar regarding the new times - will followup to that email after this meeting 16:46:58 I think it is better the subject matter expert presents - rather than just one presenter 16:47:25 rkukura: Totally agree....i want to have it as a discussion for which we need to brainstorm b4...otherwise, it may become a presentation :) 16:47:49 rkukura: Sure, thanks...in which case sukhdev, will update with the new timings 16:48:32 Would be nice for each session to start with an intro that lists the topics and how they are interelated, has the discussion on the topics, then has a couple minutes at the end to see how things all fit together, get people signed up to do the work, identify dependencies/issues, etc. 16:48:33 padkrish: my email sukhdev@arista.com 16:48:59 #sukhdev#: Thanks 16:49:17 Anything else on the design summit before we move on to bugs, reviews, etc.? 16:49:56 #topic Bugs 16:50:07 #link https://bugs.launchpad.net/neutron?field.searchtext=ml2 16:50:39 My port binding outside transactions but is at the top of the list 16:51:02 Got some good review feedback, will hopefully get an update pushed before the summit 16:51:40 Is anyone aware of any bugs that need to be dicussed, need assignees, etc? 16:52:55 #topic Spec reviews 16:53:13 I’m way behind on this myself 16:53:26 Any spec reviews anyone would like to discuss? 16:54:03 #topic Code reviews 16:54:28 we posted couple for ML2 and got comment and reply to the comment but don't know what know what is next? 16:55:24 nlahouti_: Wondering if it would help to devote good chunk of the ML2 meetings to spec reviews 16:55:40 rkukura: sorry - was pulled into another thing. I put my comment on https://review.openstack.org/#/c/91811/ 16:55:49 rkukura: That would help I believe 16:55:57 We can list the ML2-related spec reviews on the agenda, and raise awareness, get reviews signed up, ... 16:56:18 I have a question related to review https://review.openstack.org/#/c/74464/ 16:56:54 irenab: what’s the question? 16:57:04 the spec is approved, by patch is still with -2, any advice how to proceed? 16:57:50 irenab: Have you pinged markmcclain on IRC and/or email? 16:58:27 rkukura: meanwile via the patch review only, will follow up with email/irc 16:58:33 I think our time focus post-summit needs to be on working through the ML2-related spec reviews 16:58:57 and the spec reviewers should then follow through on the code reviews 16:59:07 #topic Open Discussion 16:59:15 we’ve only got a minute 16:59:20 anything anyone? 16:59:54 thanks everyone! 17:00:01 thanks 17:00:09 #endmeeting