15:00:29 #startmeeting neutron_l3 15:00:30 hi 15:00:30 Meeting started Thu Jun 4 15:00:29 2015 UTC and is due to finish in 60 minutes. The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:33 The meeting name has been set to 'neutron_l3' 15:00:35 pavel_bondar: john-davidge: hi 15:00:54 #topic Announcements 15:01:02 Liberty-1 is right around the corner. 15:01:20 #link https://wiki.openstack.org/wiki/Liberty_Release_Schedule 15:01:35 hi 15:01:40 The Neutron mid-cycle is coming up. In Fort Collins. Anyone planning to attend? 15:01:57 I am planning on it 15:02:01 carl_baldwin: yes 15:02:12 carl_baldwin: I am still in the "maybe" category 15:02:13 too far away for me unfortunately 15:02:15 will be there 15:02:19 :( 15:02:47 hi 15:03:02 #link https://etherpad.openstack.org/p/neutron-liberty-mid-cycle 15:03:14 We will miss those who are unable to attend. 15:03:33 Me too ;) 15:03:39 #topic Bugs. 15:04:06 I did some triage but still need to do some more. 15:04:14 Any bugs to bring up specifically? 15:05:07 I don’t have any. 15:05:28 gsagie: This is probably a good time to bring up your dvr patch 15:05:37 https://review.openstack.org/#/c/185368/ , not a bug but its the decomposition task 15:05:39 #link https://review.openstack.org/#/c/185368/ 15:06:00 Right, not a bug but I thought we’d squeeze it in due to your constraints. 15:06:16 i aligned it with everyones comments, and would like to get it merged first before moving to the next step 15:06:24 gsagie: I have not looked at it since my latest review. 15:06:46 I think the naming was the only open question for me. What names did you settle on? 15:07:19 carl_baldwin: what you suggested, however thats still mainly in comment and the file name, i think we can replace that going forward if we think of something better 15:07:45 Okay. I’ll have another look later today. 15:07:55 the next step/patch would be to remove the is_snat_host checks and align the unit tests and the creation in the agent 15:08:09 after that i will try to think about a composition model, but that will be more intrusive code wise 15:09:08 gsagie: Let’s get to the second step and see what that looks like and go from there. 15:09:28 okie, waiting your reviews then :) 15:09:33 gsagie: Any other open issues on this? 15:09:45 nope, and thanks for letting me go first, have to run, bye! 15:10:04 gsagie: bye 15:11:42 #topic bgp-dynamic-routing 15:11:59 tidwellr: vikram___: devvesa: hi 15:12:03 hi 15:12:04 hi 15:12:22 Rebasing for CLI and driver patch done! 15:12:50 vikram___: Excellent, do you have links handy? 15:12:52 Good news we have a Horizon expert Vish willing to help us out for horizon 15:13:29 I have over-written the same patch 15:13:40 vikram___, Thanks for introducing me as an expert, don't think I am an expert but am glad to help out 15:13:54 ;) 15:14:19 vishwanathj: There will be many opportunities for doing Horizon work for Neutron features. :) 15:14:25 Driver: https://review.openstack.org/#/c/115938/ 15:14:34 that will be awesome, appreciate the opportunity 15:14:39 CLI: https://review.openstack.org/#/c/111318/ 15:15:08 vikram___: Thanks. 15:15:19 Anything needing discussion now? 15:15:28 Carl: Ramanjenya is also willing to help us for CLI work 15:15:30 A couple issues to raise 1). I want to tread lightly with the scope of the spec https://review.openstack.org/#/c/125401/ 2). The patches we have seem to be have been implemented as a service plugin. Is there a reason to continue down that path? 15:17:02 in terms of scope, I'm concerned that we may be biting off a little too much by support both routed floating IP and routed tenant networks 15:17:34 we can take the scope discussion to the spec, we don't need to go on about it here 15:17:55 tidwellr: I think we should go for floating ips first and be sure that gets done. The tenant networks piece will depend on address scopes which is not yet ready. 15:18:21 I think doing tenant networks will be a stretch goal. 15:18:43 Let’s take the service plugin discussion to the code review and possibly the ML. 15:18:54 carl_baldwin: good plan 15:19:46 Anything else on this topic? 15:19:56 Yes 15:20:17 What is plan about BP approval and code delivery? 15:20:19 Nice work on getting the patches up to date. Has any had much luck running them in a test environment against quagga or anything? 15:20:45 That's the next step;) 15:20:48 vikram___: It looks like the BP may need one more turn. 15:21:43 I’d like to see working code in someone’s test setup. Hopefully one we can automate in the gate before merging the code. 15:22:39 carl_baldwin: no luck yet, the API / DB layer has been interesting to say the least 15:23:42 getting closer though, I'm hoping to have some code cleaned up and running next week 15:24:10 We should make that the goal, connecting to floating ips through BGP routes rather than arp in an automated test. 15:24:21 +1 15:24:26 +1 15:24:27 tidwellr: vikram___: Thanks for the update on progress. 15:24:42 We have a lot to cover, so moving on... 15:24:47 Carl: I will be on a vacation for 2 weeks... 15:24:48 #topic neutron-ipam 15:24:52 carl_baldwin, vikram___ was mentioning that there was some specific horizon work related to subnet pool that needs to be done...is there a link that you can share that has the details 15:24:54 Will ensure i sync up with ryan 15:24:56 vikram___: ack 15:25:22 vishwanathj: There is a blueprint, I will get you the link in a bit. 15:25:26 Carl: thanks 15:25:34 ok, thanks 15:25:56 johnbelamaric: pavel_bondar: ping 15:26:03 carl_baldwin: pong 15:26:06 carl_baldwin: pong 15:26:18 pavel_bondar broke the refactor/decompose patch up some more 15:26:26 We’ve had some good progress. 15:26:27 first patch just moves code around 15:26:30 #link https://review.openstack.org/#/c/187961/ 15:26:37 pretty simple 15:26:42 Ref impl and driver loader down. 15:27:05 carl_baldwin: yes! I think the link above can merge soon too 15:27:05 and plan do some more siple slices from #link https://review.openstack.org/#/c/153236/ 15:27:22 s/siple/simple/ 15:27:33 pavel_bondar: so you are going to break it up a bit more? 15:27:37 johnbelamaric: It looks like it has seen some reviews and is a smaller bit to review. 15:28:21 I think this patch is very manageable. I will review it soon. 15:28:31 carl_baldwin: thanks 15:28:50 carl_baldwin: L1 is coming fast 15:28:52 johnbelamaric: yes, there are still some copy with small modifications code in ipam_backend_mixin and ipam_non_pluggable_backend 15:29:05 pavel_bondar: ok 15:29:07 so plan to create new review from that 15:29:17 makes sense 15:29:43 pavel_bondar: so you were able to get rid of the "associate" step - you updated the ref driver too? 15:30:08 johnbelamaric, right, it works fine without it 15:30:30 I have updated dependent patches with this changes 15:30:42 ok, good 15:31:17 pavel_bondar: so, once you split the patch again, we can push on with the reviews. ETA on that? 15:31:40 ETA for split is tomorrow 15:31:50 This sounds like it is really shaping up. I will be sure to review 187961 today while I might have you guys available to discuss. It looks simple enough to merge soon. 15:32:11 carl_baldwin: great! 15:32:12 Hopefully, that will open the door for the others to fall in line tomorrow and after. 15:32:23 :) 15:32:38 carl_baldwin: sounds good! 15:32:44 Feel free to ping me on this to keep the review cycles tight. 15:33:08 carl_baldwin, sure, thanks 15:33:30 Anything else for the meeting? 15:33:40 address scopes? 15:33:54 or you mean for IPAM? I don't think so 15:34:04 #topic Address Scopes 15:34:12 There. 15:34:14 :) 15:34:44 Are there still open questions? 15:35:11 BP looks really cool ;) 15:35:22 Carl: Nice work 15:35:51 vikram___: +1. I think we found a sweet spot. Something that can be implemented within the cycle and will provide a nice foundation. 15:36:07 +1 15:36:11 One I think on whether we should allow route leaking between address scopes within Neutron. But maybe that is a follow on 15:36:15 johnbelamaric: How do you feel about the spec? 15:36:18 yes 15:36:48 carl_baldwin: i think it is looking good too. I think we can put route leaking (ie, routing between scopes without NAT) in a later cycle 15:36:52 johnbelamaric: I was thinking not to allow leaking at first. But, I’m open to discussion. 15:37:11 carl_baldwin: +1 let's keep the scope do-able! 15:37:41 johnbelamaric: I think that VRF work in Linux may help us with that. But, it is bleeding edge work that isn’t even merged. 15:37:58 carl_baldwin: yeah, who knows when that will happen… 15:38:18 johnbelamaric: Right, I want to keep an eye on it. 15:38:29 carl_baldwin: Ok, +1 from me :) 15:38:37 Ok, I’ll try to get this BP reviewed by the drivers team and approved. 15:39:01 #topic DNS 15:39:15 I’ll give a report on mlavalle ’s behalf. 15:39:46 The BPs to get nova and neutron talking have been updated and are simpler and looking good. 15:40:06 Kiall from designate is taking over the BP to get neutron and designate talking. 15:40:28 mlavalle is beginning to write code. The momentum is looking good. 15:41:05 That’s all. 15:41:10 sounds like good news 15:41:11 #topic IPv6 15:41:27 HenryG: Are you around? 15:42:01 It was mentioned yesterday that we might want to discuss bug 1460720 15:42:01 bug 1460720 in neutron "Add API to set ipv6 gateway" [Undecided,New] https://launchpad.net/bugs/1460720 - Assigned to Abishek Subramanian (absubram) 15:42:25 There is also the testing effort (bug 1401726) 15:42:25 bug 1401726 in tempest "Tempest IPv6 scenario tests use IPv4 and floating IPs to connect and test" [High,Confirmed] https://launchpad.net/bugs/1401726 - Assigned to Sean M. Collins (scollins) 15:43:57 Anyone around? 15:44:27 * haleyb is lurking 15:44:44 Okay. Will bump it to another week or discuss offline. 15:45:22 #topic ML3 15:45:49 i only know of the first bug - ipv6_gateway api support, which i talked to abishek at summit about, seems like a good step 15:46:20 I guess maybe we were looking to talk ML3 in a later meeting. 15:46:28 #topic Open Discussion 15:46:38 Any other topics? 15:47:07 I'm planning to organize my thoughts on ML3 but haven't yet 15:47:23 Armando has given me some good input to think about 15:47:27 pcarver: Ah, you did make it. Good to see you. I wasn’t sure what your nick was. 15:47:52 yeah, sorry. Wasn't watching the whole meeting, just checking occasionally 15:48:14 pcarver: Thanks for the update. I have seen your ML post and I plan to respond as soon as I get a minute. 15:48:15 carl_baldwin: New prefix delegation patchset is up to address your review comments https://review.openstack.org/#/c/158697 15:48:36 Hopefully it captures everything. Currently evaluating what extra testing the new pd classes need 15:48:45 So I know I've got L3 routing problems, but I want to structure a writeup that's not specific to just me 15:49:02 john-davidge: ack. 15:49:18 by L3 routing problems, I mean performance specifically, high bandwidth, high PPS, low latency 15:49:43 pcarver: understood 15:52:13 I did forget to mention the “network segments” topic. I removed “routing networks” and replaced it with this. 15:52:32 I don’t know what the chances are for making Liberty with it (if any) but it is something I’m thinking about. 15:53:55 Attempt to find common ground between https://bugs.launchpad.net/neutron/+bug/1458890 and https://review.openstack.org/#/c/172244/ 15:53:55 Launchpad bug 1458890 in neutron "Add segment support to Neutron" [Undecided,Confirmed] 15:54:06 Anyway, just FYI. 15:54:23 Last call for other topics before closing the meeting 15:56:08 Bye. Thanks to everyone. 15:56:10 #endmeeting