16:00:34 #startmeeting networking_ml2 16:00:35 Meeting started Wed Mar 19 16:00:34 2014 UTC and is due to finish in 60 minutes. The chair is mestery. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:39 The meeting name has been set to 'networking_ml2' 16:01:10 #link https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_March_19.2C_2014 Agenda 16:01:25 There are no action items to review from last week per the logs. 16:01:37 #topic Migration BP 16:01:46 #link https://review.openstack.org/#/c/76533/ Migration Review 16:01:58 I still need to find some time to try this out. Has anyone else tried marun's patch out by chance? 16:02:18 Hi 16:03:11 OK, if anyone has cycles to try it out, please do. 16:03:12 I found a couple issues with it (via inspection), and marun will be updating, but I think its worth testing now. 16:03:28 rkukura: Cool, thanks! 16:03:36 Anything else on the migration? 16:04:05 #topic Bugs 16:04:18 Port binding details bugs are making their way upstream now. 16:04:37 This bug (https://bugs.launchpad.net/neutron/+bug/1276395) has one merged and one approved and on it's way upstream now. 16:04:52 Thanks rkukura for those! 16:04:59 Hopefully the gate settles down a bit and lets the second in today. 16:05:08 Thanks for the reviews, team! 16:05:28 #link https://bugs.launchpad.net/neutron/+bug/1276391 Bind Port called in transactions 16:05:37 This one is also approved waiting to merge: https://review.openstack.org/#/c/79511/ 16:05:51 Thanks again rkukura here! 16:06:06 #link https://bugs.launchpad.net/neutron/+bug/1227336 Error handling for update ops 16:06:17 This one is in review from banix: https://review.openstack.org/#/c/69792/ 16:06:19 That's the same patch, its closes the 1st bug and is partial fix for 2nd 16:06:28 rkukura: I just realized that too :) 16:06:41 Regarding this one: https://review.openstack.org/#/c/69792/ Do we want this for icehouse still? 16:07:51 rkukura raised a question about how the undo gets reflect into mechanism drivers; I think we need to discuss that as I do not see calling the pre commit again as a solution 16:07:59 probably won't make it to Icehouse 16:08:00 That's the patch to try to undo an update if any MD's postcommit raises an exception 16:08:20 Yes, I think this may be soemthing to discuss in Atlanta at an ML2 session. 16:08:28 Should we officially defer this to Juno then? 16:08:34 I'm wondering if the current MDs handle this OK via re-synching or anything? 16:09:14 rkukura: I looked at a few and I think the answer is no 16:09:22 Sorry for being late 16:10:17 I think this one warrants more discussion perhaps, so should we move this to Juno then? 16:10:25 +1 16:10:27 Fine with me 16:10:27 * mestery is inclined to think that given where we're at in Icehouse. 16:10:36 yes 16:10:46 #action mestery to officially move https://launchpad.net/bugs/1227336 to Juno 16:11:07 Next bug on the list is VIF security: https://bugs.launchpad.net/nova/+bug/1112912 16:11:10 banix: I'm also a bit worried about how this would interact with the patch I'm working on to move the binding outside a transaction, which adds a 2nd transaction to commit the binding result. Would we be able to undo this whole thing? 16:12:36 rkukura, then I think we have to deal with that as well… Not sure about the implications. 16:12:38 I think nachi has updated the 1st part of this: https://review.openstack.org/#/c/67281/ 16:13:35 rkukura: Yes, looks like there are only minor nits now and a clean jenkins run and this can maybe merge. 16:13:43 Needs another core other than amotoki on it too. 16:13:55 I'm following it 16:14:25 wow just noticed the number of jobs at the gate: 70 16:14:32 rkukura: Cool, I've looked at versions of this one in the past as well. 16:14:37 banix: Yes, gate is having trouble. 16:15:04 I think the verify queue is wedged on a swift job 16:15:28 there is a small temporary fix to mlnx MD till nova parts will be available: https://review.openstack.org/#/c/81281/. Will highly appretiate review, hope it can enter Icehouse 16:16:00 irenaba: Is this related to the VIF security patch or separate? 16:16:11 separate 16:16:27 irenab: OK, we're working through the agenda on the wiki, please save updates not on there until the end. 16:16:42 sorry, thanks 16:16:49 Next item is a binding failure: https://bugs.launchpad.net/neutron/+bug/1244255 16:16:55 This one looks to be closed now last time I looked. 16:17:09 The LP bug says fix committed 16:17:21 irenab: No worries. :) 16:18:06 The Nova side of this one is still open I guess. 16:18:17 Although it says invalid. 16:18:21 * mestery is confused on the status of this one I guess. 16:19:02 rkukura: Are you reading that one the same as me? 16:19:25 mestery: I'm just as confused 16:19:52 It's not just me at least. :) 16:19:53 I'm hoping https://review.openstack.org/#/c/80829/ will help it 16:20:25 Yes 16:20:43 OK, irenab, lets discuss your bug now. :) 16:21:03 #link https://review.openstack.org/#/c/81281/ mlnx MD bug fix 16:21:16 I have discessed 16:21:28 discussed it with rkukura 16:21:28 mestery: How about asking markmcclain about his updates to https://bugs.launchpad.net/neutron/+bug/1244255? 16:21:32 irenab: This one looks ok to me, thanks for adding tests! 16:21:42 its just temporary till nova parts will get in 16:21:55 mestery: thanks! 16:21:57 irenab: OK, cool. I'll re-review today. 16:22:10 rkukura: markmcclain, if you're here, any time to quickly discuss https://bugs.launchpad.net/neutron/+bug/1244255? 16:22:15 irenab: The issue I had is that binding:profile should not be modified by the plugin/driver, but given its temporary until nova does its part, I'm OK with it 16:22:18 markmcclain: rkukura and I are confused as to it's status. :) 16:22:44 * markmcclain reads scrollback 16:23:17 ok.. so I set it to fix committed because another review had merged 16:23:21 is this bug not fixed? 16:23:55 markmcclain: That's the part we're confused about. :) 16:24:13 markmcclain: Is the fix you are referring to: https://review.openstack.org/#/c/80829/? 16:26:08 this merged: https://review.openstack.org/#/c/55000/ 16:27:52 OK, thanks for the clarification markmcclain. 16:27:54 markmcclain: That was back in January. The one I mentioned changes these again, and is approved, pending verify (which is hung on a swift job) 16:28:54 ok.. I'll keep the bug open until 80829 merges 16:29:18 * mestery ^5s markmcclain and rkukura. 16:29:20 Thanks! 16:29:44 OK, any other ML2 related bugs to discuss today which were not on the agenda? 16:29:59 https://review.openstack.org/#/c/78774/1 16:30:07 looking for reviewers 16:30:33 rcurran: I'll take this one up to review as well. 16:30:46 ty 16:31:40 Any other bugs? 16:32:17 #topic Open Discussion 16:32:30 I put a list of deferred MDs on the agenda, FYI. 16:32:44 Also, please note the deadline to submit Design Summit sessions is April 20. 16:33:01 Maybe like in Hong Kong we'll have a few sessions of ML2 where we can merge topics which are community ML2 focused. 16:33:08 Mestery can you track the type driver refactor on the agenda: https://blueprints.launchpad.net/neutron/+spec/ml2-type-driver-refactor ? 16:33:26 #action mestery to track type driver refactor going forward https://blueprints.launchpad.net/neutron/+spec/ml2-type-driver-refactor 16:33:30 asomya: +1 16:33:30 asomya: Done, and thanks for the reminder! 16:33:35 Thanks, I'm making good progress on it 16:33:40 asomya: Awesome! 16:34:09 rkukura asomya: I think we're in agreement on the approach there, right? 16:34:10 I plan to submit a session to quickly go over all ML2 BPs that do not have their own sessions, and prioritize what gets in juno. 16:34:28 rkukura: Good idea! 16:34:41 A general question - If I were to create a multi-segmented network which uses (VLAN and VxLAN) how can I do it? 16:34:56 * banix adding the BP for a new ml2 driver immediately :) 16:35:11 banix: :P 16:35:20 Sukhdev: Good question, we need documentation around that I believe. 16:35:32 Sukhdev: You would have to "bridge" the networks using provider networks currently I believe. 16:35:55 mestery: any pointers to the documentation? 16:35:58 The support for multi-segment gateways is something I want to add to the ODL ML2 driver in Juno. 16:36:11 Sukhdev: There is no documentation. Are you volunteering to write it? :) 16:36:24 mestery: :-) 16:36:41 But Sukhdev has brought up a good discussion point: Documentation. 16:36:48 mestery: investigating - if I am able to figure out - will docuemnt :-) 16:36:55 We as an ML2 team should look at what's there and look to fill gaps in before Icehouse ships. 16:37:00 Thanks Sukhdev! 16:37:15 #action Sukhdev to investigate documenting muilti-segment ML2 networks 16:37:36 Anything else for ML2 today? 16:39:01 OK, thanks for everyone's continued ML2 efforts! 16:39:15 Start thinking about design summit discussions, bugs, and documentation now. :) 16:39:21 And we'll be back next week! 16:39:27 #endmeeting