15:00:28 #startmeeting neutron_l3 15:00:34 Meeting started Thu Mar 26 15:00:28 2015 UTC and is due to finish in 60 minutes. The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:36 hi 15:00:39 The meeting name has been set to 'neutron_l3' 15:00:57 hi 15:01:02 #topic Announcements 15:01:04 #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam 15:01:15 * carl_baldwin really needs to clean up that meeting page. 15:01:59 We have a small number of topics still in play for rc1 15:02:01 #link https://launchpad.net/neutron/+milestone/kilo-rc1 15:02:49 Rc1 will be around April 9th. 15:03:05 https://wiki.openstack.org/wiki/Kilo_Release_Schedule 15:03:11 Any other announcements? 15:04:32 #topic neutron-ipam 15:05:28 Well, as most in the meeting may know we were moved to L 15:05:56 pavel_bondar is proceeding in any case to see if we can clean up the test failures 15:06:31 then, we must decide whether to try to merge that version or proceed with the larger refactoring (which will touch plugins) 15:07:16 I would like to target L1 if at all possible 15:07:25 to cut off any chance of being pushed late in the L cycle 15:07:46 johnbelamaric: That is wise. 15:08:57 I plan to keep time dedicated to reviewing and helping out where I can. 15:09:12 carl_baldwin: thank you, that will help 15:09:48 johnbelamaric, Do you have anything to discuss here? 15:10:00 carl_baldwin: no, I think we can move on 15:10:29 johnbelamaric, Thanks. I will catch up on the email thread today. 15:11:23 rtidwell is not here I don't think. 15:11:36 * carl_baldwin got his nick backwards 15:11:39 I'm here 15:11:41 tidwellr, are you around? 15:11:57 tidwellr, sorry, I do that a lot with you nick for some reason. 15:12:16 Sorry, laptop was acting up, had to log in on my phone 15:12:22 I saw that the tests are looking much better on https://review.openstack.org/#/c/157597/ 15:12:43 hi, sorry for missing meeting start 15:12:56 pavel_bondar, hi and welcome. 15:13:37 Yes, I think we should discuss quotas, Salvatore had some reasonable concerns 15:14:21 salv-orlando, are you around to discuss quotas? 15:14:32 yes I am 15:14:38 ...for 15 minutes 15:14:45 I know it would be a bit of a bummer, but I wonder if we can get by without subnetpool quotas 15:15:04 I think it might acceptable for kilo. 15:15:16 or we can just try and have "global" per-tenant quotas 15:15:27 which would acceptable and better than what we have now at the moment 15:15:38 where what we have now == nothing at all 15:15:43 salv-orlando, You mean the same quota for each tenant? 15:16:00 carl_baldwin: nope, each tenant has its own quota, but the quota is the same across pools 15:16:20 basically we should be able to do this adding pool_units as a "countableresource" in the quota API 15:17:29 but I don't know what are the requirements about quotas for subnet pools allocations from the stakeholder 15:17:47 salv-orlando, By "same across pools" mean that a tenant has only one quota to use across all pools or that the quota is just the same across pools but is counted independently? 15:18:16 carl_baldwin: good question. The current quota API has only a global counter. So it would be total number of units across pools 15:18:29 IOW, if I, a tenant, consume from one pool, does that decrease my quota for all pools? 15:18:30 the quota engine is fairly limited I'd say 15:18:44 yes, it's like with ports today 15:18:54 you have for instance 1000 ports across all networks 15:18:59 not 1000 ports per network 15:19:08 salv-orlando, okay. I understand. 15:19:29 Doesn't seem very useful for what we want/need 15:19:37 ok let's keep it as it then 15:19:54 let's say that this is an exception, then? And quotas for subnet pools are managed in their own way? 15:20:04 salv-orlando, that will be difficult. Some pools may hold very valuable ipv4 global address space. Other pools may have very cheap address space. 15:20:37 salv-orlando, Yes, that was our thinking. Sorry if that was not communicated well. 15:20:40 carl_baldwin, tidwellr: then let's forget about using the quota API. At this stage we should then approve tidwellr patch 15:21:08 if in the future we find a better way to manage quotas, we'll go back to the subnetpools API 15:22:47 salv-orlando, That sounds good to me. I did try to find a way to fit with the current API. It just wasn't going to support what we need. 15:23:11 ok. 15:23:43 I think tidwellr has only to sort out nits and then that patch can go in. 15:23:44 salv-orlando, Any other concerns about the patch or should we continue to address them in the patch? 15:23:51 carl_baldwin: I can ready your mind. 15:23:56 * read 15:24:13 Or you read and type incredibly quickly. Either way, I'm impressed. 15:24:32 salv-orlando, Thank you for your time. 15:25:12 carl_baldwin, tidwellr: np. Sorry for wasting your time by raising these concerns. 15:25:36 I do not see it as a waste of time. 15:25:39 No worries, let's do the right thing 15:25:49 tidwellr, Anything else to discuss? 15:25:58 Nope 15:27:21 tidwellr, salv-orlando: Thank you. 15:27:40 I'll link the quota patch. 15:27:43 #link https://review.openstack.org/#/c/165264 15:28:59 #topic neutron-ovs-dvr 15:29:12 I don't know if anyone is around for this. 15:29:41 Moving the dvr job to voting was discussed in the most recent Neutron meeting. 15:30:28 That is good news. 15:30:46 I thought there was a review to turn it to voting. But, I can't find it. 15:32:24 #topic Open Discussion 15:35:03 Thanks for your time everyone! 15:35:10 bye 15:35:25 #endmeeting