15:01:20 <carl_baldwin> #startmeeting neutron_l3
15:01:24 <openstack> Meeting started Thu Mar 12 15:01:20 2015 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:27 <carl_baldwin> #topic Announcements
15:01:28 <openstack> The meeting name has been set to 'neutron_l3'
15:01:47 <carl_baldwin> One more week until K3.  I can’t believe it myself.
15:01:55 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
15:02:07 <carl_baldwin> #link https://wiki.openstack.org/wiki/Kilo_Release_Schedule
15:02:35 <carl_baldwin> Any other announcements?
15:04:10 <carl_baldwin> I think we’ll have a short meeting today.
15:04:12 <carl_baldwin> #topic Bugs
15:04:21 <carl_baldwin> Any bugs we need to be aware of?
15:04:39 <pc_m> I had questions on https://bugs.launchpad.net/neutron/+bug/1431077
15:04:40 <openstack> Launchpad bug 1431077 in neutron "TRACE: attribute error when trying to fetch the router.snat_namespace.name" [Undecided,New]
15:05:23 <pc_m> When _process_added_router() called for DVR , should there be a namespace set up already?
15:05:54 * carl_baldwin looks at code...
15:07:27 <carl_baldwin> At the point _process_added_router is called, there may not be a namespace.  It is part of _router_added to create the namespace.
15:07:37 <carl_baldwin> pc_m: ^
15:08:07 <pc_m> carl_baldwin: OK, so looks like VPN code is expecting there to be a namespace.
15:09:06 <pc_m> do we know if the sequence there has changed a bit?
15:09:23 <pc_m> (did it create namespace earlier, before?)
15:09:52 <carl_baldwin> pc_m: I don’t think so but we can look at it a bit later on in the day.
15:10:09 <pc_m> carl_baldwin: sure
15:10:41 <mlavalle> pc_m: the way I understand this is that the sequence didn't change. What I discovered when working in the refactoring is that the namespace is created implcitely under certain circumstances.
15:10:46 <carl_baldwin> pc_m: I will be on-and-off a bit this morning but will be working late.
15:11:19 <pc_m> Looks like VPN expects the router to have a namespace.
15:11:42 <pc_m> I'll look at the VPN code more and can touch base later.
15:11:45 <carl_baldwin> pc_m: It seems like a reasonable expectation for vpn.
15:12:06 <carl_baldwin> pc_m: I mean, if it isn’t there vpnaas can’t do much.  :)
15:12:09 <mlavalle> pc_m: I can share with you whant I discovered after the meting and maybe we can come to a better understanding
15:12:28 <pc_m> mlavalle: that would be nice. Thanks!
15:12:49 <carl_baldwin> pc_m: mlavalle: I should be on solidly after about 11MDT and until 7MDT.  Grab me then.
15:13:00 <mlavalle> will do
15:13:10 <pc_m> roger that
15:13:17 <carl_baldwin> Any other bugs?
15:15:00 <carl_baldwin> #topic L3 Agent Restructuring
15:15:22 <carl_baldwin> I did put up one more patch to clean up some remaining router stuff in the agent.
15:16:07 <carl_baldwin> #link https://review.openstack.org/#/c/163222/
15:16:33 <carl_baldwin> It changes the sequence of some things wrt to ha routers but I didn’t think it would be a problem.
15:16:42 <carl_baldwin> amuller: ^
15:16:52 <amuller> ack
15:17:36 <carl_baldwin> It was very tangly.  :)  Teasing it apart was an interesting mental exercise.
15:18:30 <carl_baldwin> That is about all I have.  Just hoping it can all get in.
15:18:57 <carl_baldwin> #topic neutron-ipam
15:19:34 <carl_baldwin> I think we’re missing John, Salvatore, and Ryan today.
15:20:01 <seizadi> carl:john is on PTO I'm sitting in for him
15:20:20 <seizadi> I don't have anything new to report
15:20:27 <carl_baldwin> We are planning to get together tomorrow morning to discuss testing of the reference implementation integration with external IPAM.
15:20:56 <seizadi> On IRC?
15:21:15 <carl_baldwin> We don’t know yet exactly whre we’ll meet.  Possibly IRC.  If you’re interested, contact me directly and I’ll let you know where/how we will discuss.
15:21:29 <seizadi> OK, I'll
15:21:54 <carl_baldwin> It looks like the time will be 8am PDT to accomodate various parts of the world.
15:22:51 <carl_baldwin> seizadi: Thanks.
15:23:12 <carl_baldwin> Anything else?
15:23:41 <carl_baldwin> #topic dvr
15:24:12 <carl_baldwin> I don’t think I have anything to discuss and I don’t see dvr team around.
15:24:55 <carl_baldwin> #topic Open Discussion
15:26:15 <pc_m> carl_baldwin: Can you review https://review.openstack.org/#/c/162840/? I think it may avert the issue we see with the namespace.
15:26:24 <pc_m> carl_baldwin: It's a VPN commit.
15:27:01 <carl_baldwin> pc_m: Yes, I will review it.
15:27:35 <pc_m> carl_baldwin: Essentially, as part of fixing another issue, it changes the sync in VPN to not create a process manager for the VPN process and then turn around an immediately delete it, because no VPN associated with router.
15:27:45 <pc_m> Doing that, it will never ask for namespace.
15:28:09 <pc_m> Although we do need to understand why the expected namespace is not available.
15:29:04 <haleyb> pc_m, carl_baldwin: did we need a solution to force an iptables apply? i am still struggling through https://review.openstack.org/#/c/158863/ but there's a dependent VPN commit that seems to need the funcionality
15:29:47 <haleyb> it would be easy to add a flag to apply() in the short term
15:30:18 <carl_baldwin> pc_m: We can still discuss later.  I’ll take a look at the review.
15:30:33 <pc_m> carl_baldwin: OK.
15:30:39 <carl_baldwin> haleyb: What is the dependent VPN commit?
15:30:56 <haleyb> https://review.openstack.org/#/c/158865/
15:30:59 <pc_m> haleyb: So this is to defer the apply() and VPN applies immediately?
15:31:28 <pc_m> that's FWaaS
15:31:49 <haleyb> doh, you're right
15:32:09 <pc_m> had me confused a bit as I didn't know of such change.
15:32:12 <carl_baldwin> haleyb: pc_m:  Makes more sense now.
15:32:13 <carl_baldwin> ;)
15:32:29 <haleyb> yes, i will go get more coffee :)
15:33:34 <haleyb> carl_baldwin: you can ping me offline on that if you want
15:33:58 <carl_baldwin> haleyb: ok.  Let’s chat about it a bit later.
15:36:40 <carl_baldwin> Last call for Open Discussion...
15:36:43 <carl_baldwin> :)
15:37:13 <carl_baldwin> Thanks everyone.
15:37:20 <carl_baldwin> #endmeeting