14:02:11 #startmeeting networking_ml2 14:02:12 Meeting started Wed Oct 9 14:02:11 2013 UTC and is due to finish in 60 minutes. The chair is rkukura. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:02:12 rkukura: Error: "startmeeting" is not a valid command. 14:02:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:02:15 The meeting name has been set to 'networking_ml2' 14:03:12 #link https://wiki.openstack.org/wiki/Meetings/ML2 Agenda 14:03:13 rkukura: Error: "link" is not a valid command. 14:04:09 mestery is not available today, so I'll chair today's meeting 14:04:27 not sure if I should worry about the uvirtbot errors? 14:04:46 good morning 14:05:13 hi 14:05:18 hi 14:05:39 I think today's meeting will be quick 14:05:56 Three items: RC2, docs, and summit 14:06:14 #topic RC2 14:06:16 rkukura: Error: "topic" is not a valid command. 14:06:37 markmcclain is planning to cut RC2 tomorrow 14:07:08 yes 14:07:16 One ML2 bug still in review and targeted for RC2 is https://review.openstack.org/#/c/49644/ 14:07:21 is there any blocker issue? 14:07:42 This is the 2nd part of fixing auto-deletion 14:08:11 so, this is a similar fix to network deletion, right 14:08:16 it was supposed to just do for delete_subnet what was already done for delete_network 14:08:24 I've added 1234195 to RC2 14:08:51 but the review is exposing issues with DB locking 14:09:16 we should consider this https://bugs.launchpad.net/neutron/+bug/1237082 for RC2 14:09:17 Launchpad bug 1237082 in neutron "ml2 vxlan interface with linuxbridge use proxy mode" [Undecided,In progress] 14:10:18 vxlan without l2-pop is not functionnal anymore 14:10:24 on 1234195, I'm leaning towards not trying to add the "with_lockmode('update') calls, since they don't work with outer joins on postgress 14:10:26 with linuxbridge 14:10:38 and they don't necessarily prevent new rows from being added 14:10:53 question is - should we try to do some sort of table-level locking? 14:10:58 or leave this for later? 14:12:28 matrohon: 1237082 does have the havnana-rc-potential tag 14:13:27 rkukura : i'v put the tag, hope it will be considered for rc2 14:14:38 This looks straightforward, but may need a unit test 14:15:01 Lets try to get this in today 14:15:20 Any other patches or bugs that should be considered for RC2? 14:15:31 it is written and will be pushed as soon as tox have finished 14:15:50 matrohon: Great! I'll review it ASAP. 14:16:29 it will be toot hsort for rc2, but live-migration with l2-pop doesn't work 14:16:43 i will work on it asap 14:17:10 ok, anything else? 14:17:21 matrohon: we can target that for fix for the first stable update 14:17:34 markmclain: great 14:17:48 #topic: Docs 14:17:50 rkukura: Error: "topic:" is not a valid command. 14:18:15 The l2-pop mechanism driver uses an ARP responder to avoid ARP broadcast but it only works with the LB agent. We need to implement an ARP responder for the OVS agent also. 14:18:35 I've open a bug https://bugs.launchpad.net/neutron/+bug/1237427 14:18:37 Launchpad bug 1237427 in neutron "l2-pop ML2 mechanism driver need ARP responder with the OVS agent" [Undecided,New] 14:18:39 I know emagana got basic stuff on ML2 into the cloud admin guide 14:18:49 An ebtables manager could be able to do that. I've started a patch for that. 14:18:52 It's not too intruisive for the H release because it's only impact code if the l2-pop mechanism driver is enabled (that's not the case by default). 14:19:23 we can followup with more detail in the Cloud Admin Guide if needed 14:19:32 Do you think it could be implemented for the H release ? 14:20:26 On the subject of Docs, I have update/issue regarding https://bugs.launchpad.net/openstack-manuals/+bug/1230276 14:20:27 Launchpad bug 1230276 in openstack-manuals "Add neutron ml2 plugin info to installation guide" [High,Confirmed] 14:20:27 doude: no 14:20:56 doude: we've basically run out of time, so we'll fix in Icehouse and then backport to Havana.1 release 14:21:31 markmcclain: ok 14:21:48 Sukhdev: go ahead 14:22:12 I pulled the latest information from Github regarding install-guide and built is, it shows Neutron Chapter empty with - FIXIT in it 14:22:31 so, it seems like there is no chapter on Neutron in Intall-Guide, 14:22:41 is this true? 14:23:40 I see Neutron in all other guides, but, not in istall-guide 14:23:58 annegentle_: Is neutron info in the install guide, or being added? 14:24:32 rkukura: we are hoping you have people who are interested in adding it, testing it. The basic-install-files directory contains the grizzly release instructions, which could be updated and added 14:24:57 rkukura: here's what's still needed http://lists.openstack.org/pipermail/openstack-docs/2013-October/002936.html 14:25:13 rkukura: Nermina Miller was going to reach out to you all to see if you have content to add 14:25:37 Sukhdev: Can you work with Nermina on this? 14:26:47 roaet: were you working on the config guide? 14:26:49 rkukura: can you give me their contact info and I will reach out 14:27:32 Sukhdev: Sure, should be in the previous email thread regarding the ML2 docs tasks, but I'll find it and resend 14:27:44 Ah OK - thanks 14:28:37 BTW, I started to work onhttps://wiki.openstack.org/wiki/Neutron and noticed that under current project status, it talks about Folsom 14:28:59 So, this seems to be out of date 14:30:05 I added a section about Havana and ML2 update but did not touch other sections https://wiki.openstack.org/wiki/Neutron 14:30:33 markmcclain: Do you know if the main neutron wiki page is slated for updating/cleanup? 14:30:58 rkukura: it needs a clean-up 14:31:17 I was hoping coordinate work on it after Havana releasae 14:31:23 Sukhdev: The new material looks good 14:31:51 Thanks....I was tempted to do it, but, was not sure if I mess with it 14:31:55 markmcclain: thanks 14:32:01 I mean clean up 14:32:21 Sukhdev: I think any of use can/should cleanup this sort of thing if we have the opportunity 14:32:52 Is roaet around? 14:33:21 I can remove the current project status and spread the information around appropriately - 14:33:41 Last I had from him on the config guide was regarding the mechanism being used to extract info from the code 14:35:23 We'll follow up on config guide later, but I think this is really important to have complete when havana ships 14:35:48 anything else on docs? 14:36:32 #topic Icehouse Design Summit topics 14:36:34 rkukura: Error: "topic" is not a valid command. 14:37:25 We brainstormed some candidates last week. 14:37:59 markmcclain said the cutoff to submit is 10/17, but I don't think we should wait til the last moment 14:38:30 please don't :) 14:38:58 I'm thinking we should go ahead and submit our ideas rather than trying to pre-coordinate them into a small number of ml2 topics 14:39:50 Any objections to that plan? 14:41:26 Next week we can look at what has been submitted, see if anything is missing, and maybe come up with suggestions for markmcclain on what can be combined 14:42:09 anything else on the design summit? 14:42:10 is there a wiki listing what has already been propose/planned? 14:42:24 http://summit.openstack.org/ 14:43:13 the minutes from last weeks ml2 meeting are at http://eavesdrop.openstack.org/meetings/networking_ml2/2013/ 14:43:32 #topic Open Discussion 14:43:33 rkukura: Error: "topic" is not a valid command. 14:43:56 Anything else to discuss today? 14:44:12 do we consider having an ml2 agent with drivers for lb , ovs and others? 14:44:36 it is not listed in the last week minute 14:44:45 matrohon: I think that would make sense to propose 14:45:22 "14:39:23 modular agent is another" 14:46:18 matrohon: There were a lot of suggestions in the log that didn't make the minutes 14:46:24 Anything else for today? 14:47:04 Thanks everyone! 14:47:12 thanks 14:47:18 tahnks 14:47:23 #endmeeting