15:02:39 <carl_baldwin> #startmeeting neutron_l3
15:02:40 <openstack> Meeting started Thu Jun 18 15:02:39 2015 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:02:43 <openstack> The meeting name has been set to 'neutron_l3'
15:02:51 <regXboi> #info regXboi (Ryan Moats)
15:02:58 <carl_baldwin> #topic Announcements
15:03:02 <regXboi> carl_baldwin: as promised/threatened :)
15:03:25 <ramanjaneya_> Hi carl
15:03:35 <haleyb> hi
15:03:46 <yamahata> hello
15:03:53 <carl_baldwin> I’d like to welcome haleyb as our first core focusing on the L3 area.
15:03:53 <yamamoto> hi
15:03:54 <johnbelamaric> hi
15:04:01 <pavel_bondar> hi
15:04:05 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
15:04:08 <pc_m> Yay!
15:04:13 <yamamoto> haleyb: congrats
15:04:39 <johnbelamaric> haleyb: welcome, congrats
15:04:39 <mlavalle> haleyb: yeah, congratulations!
15:04:41 <carl_baldwin> The Neutron mid-cycle is next week in Fort Collins.  I look forward to seeing as many of you as possible.
15:05:03 <johnbelamaric> carl_baldwin: looks like I can make it now :)
15:05:16 <carl_baldwin> johnbelamaric: great
15:05:21 <carl_baldwin> #link https://etherpad.openstack.org/p/neutron-liberty-mid-cycle
15:05:24 <haleyb> thanks, i look forward to helping out more
15:05:29 <carl_baldwin> Also, Liberty-1 is next week too.
15:05:41 <regXboi> carl_baldwin: is there any change the mid-cycle will have an IRC channel or way for remotes to get involved?
15:05:50 <regXboi> s/change/chance/
15:05:57 <mohankumar> Hi
15:06:21 <carl_baldwin> regXboi: mestery has put a note at the bottom of the etherpad.
15:06:33 <carl_baldwin> #topic Bugs
15:06:43 <regXboi> carl_baldwin: thanx, I see that - will bug mestery (I'm good at that)
15:07:43 <carl_baldwin> Thanks to mlavalle for helping us to stay on top of the bugs with the l3-ipam-dhcp tag
15:07:53 <mlavalle> carl_baldwin: :-)
15:08:01 <carl_baldwin> Our bug list is shrinking.  That is good.  Just in time to write some more.  ;)
15:08:27 <carl_baldwin> #topic Routing Network Segments
15:08:39 <carl_baldwin> This is the hot topic for today, I think.
15:08:43 <regXboi> :)
15:09:22 <carl_baldwin> So, where do we start?  I guess I could get a couple of links.
15:09:42 <regXboi> always good for the minutes :)
15:10:12 <HenryG> I am a bit out of the loop on this. Is it intended to cover all the various segmentation strategies that the operators are mentioning (no two seem to be the same)?
15:10:12 <carl_baldwin> #link https://bugs.launchpad.net/neutron/+bug/1458890
15:10:14 <openstack> Launchpad bug 1458890 in neutron "Add segment support to Neutron" [Undecided,Triaged]
15:10:18 <carl_baldwin> ^ this is the rfe
15:10:47 <carl_baldwin> HenryG: I guess that is what we’re here to discuss.
15:11:01 <carl_baldwin> #link https://review.openstack.org/#/c/172244/
15:11:14 <carl_baldwin> ^ This is the spec that I wrote with a similar network segments concept.
15:11:47 <regXboi> I guess the first question is whether the spec covres the rfe?
15:12:44 * HenryG pings baoli to pay attention because he was asking about this
15:12:51 <carl_baldwin> regXboi: Not entirely.  The spec is written for a different use case.  However, it does introduce network segments which map to a given set of hosts.
15:13:14 <carl_baldwin> I believe that the network segment idea could be shared between the two.
15:13:22 <regXboi> carl_baldwin: I think I saw a comment from you that the spec was at best partial coverage because of that (am I remembering correctly)?
15:13:46 <carl_baldwin> Beyond network segments, the rfe talks about some Nova work to pass a segment id when creating a port and then using IP usage information in scheduling.
15:14:25 <carl_baldwin> regXboi: Yes, just to be clear:  The spec does not completely cover the use case in the rfe.
15:14:44 <regXboi> carl_baldwin: a concern - how would this rfe interact with the multiple provider extension?
15:15:07 * regXboi worries about overloading the segment concept
15:15:35 <carl_baldwin> regXboi: Good point, it did not cross my mind yet.
15:16:38 <regXboi> carl_baldwin: I would like to think that we could start from that point, although it risks conflating physical segmentation with the rfe, which wants to include logical segmenation
15:17:22 <carl_baldwin> regXboi: Right, I think it is worth some of my time to reconcile these with multiple provider extension.
15:17:39 <regXboi> carl_baldwin: I'd be willing to pitch some cycles there as well
15:18:45 <regXboi> because I think it might be possible to pull all three (mpe, rfe, backing networks) together
15:19:43 <carl_baldwin> regXboi: I think you’re right.  At least find the commonality and use it to build a good base for the rfe and routing network segments to build from.
15:19:45 <pc_m> regXboi: Any links to info on the multiple provider extension?
15:20:05 <regXboi> #link http://developer.openstack.org/api-ref-networking-v2-ext.html
15:20:13 * pc_m trying to come up to speed on this
15:20:20 <pc_m> regXboi: Thanks!
15:20:25 <regXboi> pc_m: yw
15:20:37 <carl_baldwin> pc_m: Under “Networks multiple provider extension”
15:20:59 <pc_m> yeah found it. thanks.
15:21:00 <carl_baldwin> regXboi: Let’s do some homework today and maybe we could meet in the neutron room tomorrow at some time for a discussion.
15:21:31 <regXboi> carl_baldwin: sounds like a plan - I'll check my calendar and hit you with some times I can be there
15:22:29 <carl_baldwin> regXboi: Great.  Anyone else want to be notified of the time?
15:22:40 <carl_baldwin> If so, ping regXboi
15:22:55 <pc_m> sure
15:22:56 <carl_baldwin> Okay.  Anything else on network segments that we should hit today?
15:24:07 <carl_baldwin> I plan to attend the LDT meeting after this meeting.  We may mention it there too.
15:24:38 <regXboi> carl_baldwin: feel free to let folks there know to ping me for notification about tomorrow's times
15:24:46 <carl_baldwin> regXboi: I will.
15:24:50 <regXboi> :)
15:24:52 <pc_m> carl_baldwin: What channel?
15:25:06 <carl_baldwin> pc_m: LDT?
15:25:09 <pc_m> for the LDT meeting
15:25:17 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/LDT
15:25:30 <pc_m> thx
15:25:56 <carl_baldwin> #topic bgp-dynamic-routing
15:26:13 <carl_baldwin> tidwellr: ping
15:26:21 <tidwellr> hey
15:26:35 <carl_baldwin> How are we doing here?
15:27:09 <devvesa> Hi. I think I log in time
15:27:39 <carl_baldwin> devvesa: You are here.  Hi.  Didn’t mean to miss you.
15:27:52 <tidwellr> I think we're in good shape .  I'm working on another patch set for the API/DB layer. I'm taking the time to do API tests as I go.
15:27:56 <devvesa> Oh, just logged now :)
15:28:03 <devvesa> Sorry
15:28:38 <carl_baldwin> tidwellr: Are all of the patches on a topic we could like to here?
15:28:48 <carl_baldwin> devvesa: Good timing.
15:29:12 <tidwellr> bp/bgp-dynamic-routing
15:29:59 <carl_baldwin> #link https://review.openstack.org/#/q/topic:bp/bgp-dynamic-routing,n,z
15:30:10 <regXboi> tidwellr, carl_baldwin: is there a plan for incorporating these tests into the gate?
15:30:27 <regXboi> I ask because I suspect the LDR folks will want to tune that test for their specific deployments
15:30:37 <tidwellr> yes,  that needs to be done, not sure how to start though
15:31:02 <regXboi> (sorry LDT, not LDR) or is the plan to have that in 3P/CI testing so that each LDT can tune it as needed?
15:31:51 <tidwellr> I've a little myopic about just getting tests running in my environment, but my intention is to have a CI job for BGP
15:32:29 <tidwellr> could use a little help there......
15:33:21 <regXboi> tidwellr: unfortunately, I'm not sure I can commit to that, but let me look around and see if I can scare up some help...
15:34:33 <carl_baldwin> Well, let’s hit the reviews and keep our eye on building tests for the gate.
15:34:38 <carl_baldwin> Anything else for this topic?
15:34:59 <carl_baldwin> #topic neutron-ipam
15:35:17 <pavel_bondar> hi
15:35:19 <tidwellr> not from me, devvesa?
15:36:02 <carl_baldwin> pavel_bondar: great work on breaking up the large review.  It is a lot easier to review now.  It is nice to have check-points along the way.
15:36:23 <devvesa> neither. Just waiting your next patch, tidwellr :)
15:36:24 <pavel_bondar> I have created 3 more review today:)
15:36:38 <pavel_bondar> next review in chain to be merged is #link https://review.openstack.org/#/c/189299/
15:37:08 <pavel_bondar> and 6 more reviews are on top of that
15:37:46 <johnbelamaric> pavel_bondar: nice, lots of work but I think it is much easier this way
15:38:18 <carl_baldwin> pavel_bondar: This one is looking pretty good.
15:38:21 <pavel_bondar> johnbelamaric: agree
15:38:38 <pavel_bondar> carl_baldwin: nice
15:39:12 <pavel_bondar> I can post full review chain into IRC if needed
15:39:15 <carl_baldwin> So, we just keep going.  Reviewers can start with this^ and follow the chain.
15:39:37 <carl_baldwin> pavel_bondar: Probably not needed.
15:39:49 <pavel_bondar> carl_baldwin: ok
15:39:58 <carl_baldwin> Anything to discuss?
15:40:21 <carl_baldwin> #topic ML3 Router plugin
15:40:31 <carl_baldwin> Anyone here to give a report on this?
15:40:37 <yamahata> Hello.
15:40:46 <yamahata> So far we gathered the use cases in etherpad
15:40:53 <yamahata> https://etherpad.openstack.org/p/neutron-modular-l3-router-plugin-use-cases
15:41:17 <yamahata> They are for requirement.
15:41:43 <yamahata> So the next step is to define operation scenario in order to define desired API or CLI operations
15:42:00 <yamahata> pcarver: ping?
15:42:01 <regXboi> yamahata: Use Case #4 looks very close to service chaining - is that intentional?
15:42:08 <carl_baldwin> These look good.
15:42:34 <yamahata> regXboi: Yeah. the difference is that the service is on edge.
15:42:46 <yamahata> If SFC supports it, it's quite fine with it.
15:43:52 <yamahata> Deployer may think SFC is overkill. they may want more simpler way to just deploy single function.
15:43:52 <regXboi> yamahata: got it - but I need to go off and think about whether you've just created an SFC implementation unintenionally
15:44:12 <regXboi> yamahata: I'm not saying that's a bad thing
15:44:24 * regXboi needs to think about the implications
15:44:40 <yamahata> regXboi: I don't want to reinvent SFC.
15:45:02 <yamahata> pcarver doesn't seem here. Contact him off line.
15:45:18 <regXboi> yamahata: understood, like I said, let me think about the implications
15:45:31 <yamahata> regXboi: sure.
15:45:45 <carl_baldwin> yamahata: How do we go about the next step?
15:46:13 <yamahata> I'll write up a draft for desired operation scenario. then get review.
15:46:27 <carl_baldwin> yamahata: Sounds like a plan.
15:46:43 <carl_baldwin> Anything else?
15:46:55 <carl_baldwin> #topic dns-resolution
15:46:58 <carl_baldwin> mlavalle: ping
15:47:25 <mlavalle> Kiall helped us to respond to the concerns in https://review.openstack.org/#/c/88623
15:47:40 <carl_baldwin> I was looking at those this morning.
15:47:52 <mlavalle> He added a comment yesterday. I'll update the spec today
15:48:29 <mlavalle> He also wrote the discussion in vancouver to https://review.openstack.org/#/c/88624/, so it is looking good
15:49:26 <mlavalle> The nova side https://review.openstack.org/#/c/90150/ was refreshed by me laste last week andgot a +1. Will follow up with johnthetubaguy to get it reviewed by him
15:49:53 <regXboi> mlavalle: I assume you are looking for reviews on updated 88623 and 88624?
15:50:03 <mlavalle> And I am already coding for https://review.openstack.org/#/c/88623
15:50:17 <mlavalle> regXboi: yes
15:50:18 <carl_baldwin> mlavalle: Good progress.
15:50:33 <regXboi> mlavalle: thx
15:50:36 <mlavalle> I should push WIP patchset soon
15:51:19 <mlavalle> Kial comments to 88623 change code a bit, but not much.
15:51:27 <mlavalle> that's all I have
15:51:51 <carl_baldwin> mlavalle: I look forward to seeing it.  Ping at least haleyb and me for reviews.  And, anyone else interested.
15:52:00 <mlavalle> will do
15:52:23 <carl_baldwin> #topic Address Scopes
15:52:28 * mlavalle looking forward to get review from our brand new core :-)
15:52:54 <carl_baldwin> I added an update to the BP:  https://review.openstack.org/#/c/192914/
15:53:24 <carl_baldwin> Otherwise, I’m coding now and hope to have things working on the L3 agent side in a couple of days.
15:53:41 <carl_baldwin> The address scope crud work has also started.
15:54:02 <carl_baldwin> https://review.openstack.org/#/c/189741/
15:54:17 <carl_baldwin> Vikram is still on vacation but should return soon.
15:54:26 <carl_baldwin> #link https://review.openstack.org/#/c/189741/
15:54:44 <carl_baldwin> #topic IPv6
15:55:09 <carl_baldwin> Anything here?  I know I’ve been neglecting the PD reviews.  I’ve been trying to get to it and will soon.
15:55:40 <john-davidge> carl_baldwin: Thanks, I know it’s a big patch!
15:57:36 <carl_baldwin> In general, this needs more reviewer attention.  Not just from cores.  I’d like to see it start to build some consensus with multiple +1s.
15:58:04 * haleyb needs to re-review the PD changes as well
15:58:20 <john-davidge> Here’s the first PD patch in the chain: https://review.openstack.org/#/c/158697
15:59:05 <carl_baldwin> How can we get more attention on this from the community?
15:59:08 <john-davidge> The second is currently in merge conflict with the ongoing db_plugin_v2 decomp work
16:00:12 <john-davidge> carl_baldwin: Reviews keeping it near the top of the list will help, otherwise I could spam the mailing list with demo videos :)
16:00:13 <carl_baldwin> I guess our time is up.
16:00:25 <carl_baldwin> We can take stuff to the neutron room.
16:00:27 <carl_baldwin> #endmeeting