14:01:26 <carl_baldwin> #startmeeting neutron_routed_networks
14:01:27 <openstack> Meeting started Tue Jun 14 14:01:26 2016 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:31 <openstack> The meeting name has been set to 'neutron_routed_networks'
14:01:45 <mlavalle> o/
14:01:49 <carl_baldwin> How is everyone?
14:02:06 <carl_baldwin> #topic Announcements
14:02:38 <rtheis> hi
14:02:51 <carl_baldwin> We've got about a month to the Newton-2 milestone
14:02:53 <carl_baldwin> #link http://releases.openstack.org/newton/schedule.html
14:03:04 <carl_baldwin> Oh, and the etherpad we're using is here:
14:03:09 <carl_baldwin> #link https://etherpad.openstack.org/p/routed-provider-networks-notes
14:04:08 <carl_baldwin> Neutron sprint is Cork Ireland on August 17-19.
14:04:24 <carl_baldwin> #link https://etherpad.openstack.org/p/newton-neutron-midcycle
14:04:33 <carl_baldwin> I'm still working on getting approvel
14:04:37 <carl_baldwin> *approval
14:04:49 <john-davidge> Me too, remaining hopeful
14:05:12 <carl_baldwin> Any other announcements?
14:06:03 <carl_baldwin> #topic Development Environment
14:06:33 <carl_baldwin> mlavalle has done a lot of work on this.  I started looking at it yesterday and will try it out this morning as soon as I can.
14:06:44 <carl_baldwin> mlavalle: you want to tell us about it?
14:06:54 <mlavalle> yes!
14:07:12 <mlavalle> The team can find the Vagrant environment here:
14:07:25 <mlavalle> #link https://github.com/miguellavalle/routednetworksvagrant
14:07:51 <mlavalle> It is a 3 nodes environment: allinone, compute1 and compute2
14:08:14 <mlavalle> And 2 physical networks: physnet1 between allinone and compute 1
14:08:29 <mlavalle> and physnet2 between allinone and compute 2
14:08:44 <mlavalle> That gives us 2 'compute racks' to play with
14:08:56 <mlavalle> each rack with 2 computes
14:09:09 <mlavalle> in the near future I'll add a router VM
14:09:13 <carl_baldwin> mlavalle: are there 2 allinone nodes?
14:09:28 <mlavalle> no, only 1 allinone
14:09:49 <mlavalle> it is connected in  the 2 'racks'
14:10:02 <mlavalle> so please go play with it!
14:10:04 <carl_baldwin> mlavalle: We have an assumption currently that a compute host is only connected to one segment.
14:10:40 <mlavalle> in that case I can easily add a compute3 and we will have 2 racks that way
14:10:52 <mlavalle> not a big deal
14:10:57 <carl_baldwin> mlavalle: ok
14:11:02 <mlavalle> I will do that today
14:11:13 <carl_baldwin> mlavalle: Sounds cool.
14:11:17 <mlavalle> that's it
14:11:46 <carl_baldwin> mlavalle: Thanks!
14:12:39 <carl_baldwin> #topic Associating Subnets to Segments
14:12:56 <carl_baldwin> There is just one client patch to look at.  Looks like it needs a follow up.
14:13:05 <carl_baldwin> #link https://review.openstack.org/#/c/315573/
14:13:23 <xiaohhui> yes, will update it soon
14:13:28 <rtheis> OSC patch is here: https://review.openstack.org/#/c/328500/
14:14:05 <carl_baldwin> rtheis: Thanks.
14:14:12 <rtheis> yw
14:15:00 <carl_baldwin> So, the IPAM stuff is looking pretty good.  All of the patches have merged and IPAM is not aware of segments.
14:15:34 <carl_baldwin> #topic Deferred IP allocation in Nova
14:15:40 <carl_baldwin> This is what I'm working on currently.
14:16:09 <carl_baldwin> I put up a simple strawman patch yesterday for discussion.
14:16:12 <carl_baldwin> #link https://review.openstack.org/329221
14:17:15 <carl_baldwin> Take a look if you get a chance.  I need to catch up on feedback on it.
14:18:03 <carl_baldwin> Looks like johnthetubaguy has left me some feedback to look at.  Thanks!
14:18:37 <carl_baldwin> #topic Schedule DHCP to Segments and Create DHCP ports
14:18:45 <carl_baldwin> blogan: You around?
14:19:05 <Reedip_> Hi
14:19:58 <carl_baldwin> Maybe not.
14:20:05 <carl_baldwin> I know progress has been made on the patches.
14:20:40 <carl_baldwin> Let's review the patches listed.
14:20:48 <carl_baldwin> #link https://review.openstack.org/#/c/326261
14:20:57 <carl_baldwin> #link https://review.openstack.org/#/c/311931
14:21:09 <carl_baldwin> #topic Integration with Nova Scheduler
14:21:51 <carl_baldwin> I think this is now the long pole.
14:22:01 <carl_baldwin> mlavalle: still around?
14:22:06 <mlavalle> hi
14:22:09 <mlavalle> again :-)
14:22:47 <mlavalle> Yes, half of the focus here is the refactoring of 'allocate_for_instance' in the Neutron API within Nova
14:23:02 <mlavalle> the gerrit topic is:
14:23:11 <mlavalle> #link https://review.openstack.org/#/q/topic:bp/prep-for-network-aware-scheduling
14:23:49 <mlavalle> johnthetubaguy has been creating a series of small patchset
14:24:21 <mlavalle> They culminate (so far) with https://review.openstack.org/#/c/328966/
14:24:40 <carl_baldwin> mlavalle: Looks like the kind of series which could use some core reviewer attention to get a merge cadence going.
14:24:58 <johnthetubaguy> so it sounds like we might get blocked on get-me-a-network stuff
14:25:00 <mlavalle> This is where allocate_for_instance is split into create/get port and update port phases
14:25:05 <johnthetubaguy> but hopefully that will not slow things down too much
14:25:25 <johnthetubaguy> there will be a few more patches, before that end one will be finished, but making progress
14:26:24 <carl_baldwin> johnthetubaguy: Are there other Nova cores with bandwidth to review and merge some of the earlier ones?
14:27:07 <carl_baldwin> I know what its like to maintain a long chain of patches.
14:27:10 <johnthetubaguy> I have asked around, so hopefully that will get attention soon ish
14:27:38 <johnthetubaguy> thats when the get-me-a-network stuff came up, I need to go look at that to see how bad the clash is
14:28:12 <carl_baldwin> johnthetubaguy: thanks.
14:28:19 <carl_baldwin> johnthetubaguy: Anything we can do to help?
14:30:20 <carl_baldwin> #topic L2 Adjacency Extension
14:30:27 <carl_baldwin> Reedip_: hi
14:30:46 <Reedip_> Hi carl_baldwin
14:31:00 <carl_baldwin> Will this get an update soon?
14:31:09 <Reedip_> I was busy with some other works for the past 2 weeks, why was not able to work on it
14:31:32 <carl_baldwin> Reedip_: ok, ping me when you get a chance to update it.
14:31:35 <Reedip_> Yes , I will create the next patch tomorrow
14:31:41 <Reedip_> Yes, I would need your help
14:31:42 <carl_baldwin> #topic Enable Create / Delete of Segments on Existing Network
14:31:51 <carl_baldwin> xiaohhui: Are you around?
14:31:55 <xiaohhui> yes
14:32:08 <johnthetubaguy> carl_baldwin: sorry, got distracted, right now the reviews you folks are doing on there is great stuff
14:32:21 <xiaohhui> I got one feedback last week, and address it.
14:32:28 <xiaohhui> But still need some more reviews
14:32:41 <mlavalle> johnthetubaguy: ++
14:32:48 <carl_baldwin> johnthetubaguy: we'll keep that up then.  Thanks!
14:33:33 <carl_baldwin> xiaohhui: Thanks for your work on this.  I'll review again today.
14:33:42 <xiaohhui> OK, thanks
14:34:42 <rtheis> xiaohhui: nice work on that, I tested out the patch set and it worked well for me...no issues
14:34:55 <carl_baldwin> xiaohhui: +1
14:35:01 <xiaohhui> Thanks...
14:35:14 <xiaohhui> I also did some tests in my local env
14:35:21 <carl_baldwin> xiaohhui: Also, thanks for your work on that ML2 bug around returning fixed_ips on port update.  That merged.
14:35:42 <xiaohhui> YW
14:35:56 <carl_baldwin> #topic Client
14:36:07 <carl_baldwin> rtheis: Anything more to discuss?
14:36:43 <rtheis> I rebased the "network segment create" and "network segment delete" patches for OSC and SDK and tested with xiaohhui patch
14:36:59 <rtheis> they should be ready to go once the neutron support merges
14:37:28 <rtheis> I also wanted to ask about other client support needed.
14:37:51 <rtheis> Beyond the network segment CRUD, and associating subnets to segments...  are there any other client needs?
14:38:19 <rtheis> I'm tracking name and description for segments as one possible need
14:39:16 <Reedip_> rtheis I think that was stuck in a merge conflict , will pull up the new patch today( or maybe I already posted it??)
14:39:27 <mlavalle> rtheis: do you have your pathsets links handy? I would like to test tham in the Vagrant environment
14:39:33 <carl_baldwin> rtheis: Not sure.  DHCP work shouldn't need anything new from the client.  I can't think of anything else at the moment.  I'll let you know when I do.
14:39:41 <rtheis> thanks
14:40:01 <rtheis> mlavalle: see "Client" section in https://etherpad.openstack.org/p/routed-provider-networks-notes
14:40:11 <mlavalle> rtheis: ok, thanks
14:40:17 <rtheis> yw
14:40:28 <carl_baldwin> This actually makes me think that maybe we should start writing the docs for deploying routed networks using your client patches.  That might expose something that we've missed.
14:40:57 <rtheis> sounds good
14:41:05 <carl_baldwin> I've had the networking guide in mind a little bit this week.
14:41:10 <rtheis> do we have a docs section in etherpad?
14:41:20 <mlavalle> in fact, we that guide I can configure it in the Vagrant environment and create segments with them
14:41:22 <carl_baldwin> rtheis: No, we don't yet.
14:41:50 <carl_baldwin> Let's add a section.
14:42:54 <carl_baldwin> I just added it.
14:43:00 <carl_baldwin> L48
14:43:16 <carl_baldwin> I'll ask Sam-I-Am where to start.
14:43:42 <carl_baldwin> Anything else on the client?
14:43:59 <rtheis> carl_baldwin: nothing else
14:44:16 <carl_baldwin> rtheis: thanks!
14:44:22 <carl_baldwin> #topic OVN Plugin
14:44:30 <xiaohhui> hi
14:44:32 <carl_baldwin> xiaohhui: rtheis:  Anything to discuss here?
14:44:41 <mlavalle> hi again :-)
14:44:55 <xiaohhui> We have this depency for ovn code
14:44:56 <xiaohhui> https://review.openstack.org/#/c/322486/
14:44:59 <xiaohhui> in neutron
14:45:07 <xiaohhui> It already gets a +2
14:45:20 <carl_baldwin> xiaohhui: I can give this a review today.
14:45:28 <xiaohhui> Thanks,
14:45:41 <xiaohhui> I think we will soon have segment-host-mapping in ovn.
14:45:52 <carl_baldwin> xiaohhui: Excellent.
14:45:53 <rtheis> I'm working on some ML2 enhancements for OVN driver related to segment network type support
14:46:05 <rtheis> https://review.openstack.org/#/c/327295/
14:46:39 <rtheis> https://review.openstack.org/#/c/327278/
14:46:57 <rtheis> nothing else from me
14:47:03 <xiaohhui> This patch is to update the segment-host-mapping in ovn,  https://review.openstack.org/#/c/322488/
14:47:10 <xiaohhui> and nothing else from me
14:47:40 <carl_baldwin> rtheis: Thanks
14:48:11 <carl_baldwin> xiaohhui: Thanks.  Looks like good progress.
14:48:35 <carl_baldwin> #topic Router External Gateways on Routed Networks
14:49:07 <john-davidge> Hi
14:49:19 <carl_baldwin> For this, we really need to push the service subnets spec forward.
14:49:22 <carl_baldwin> #link https://review.openstack.org/#/c/300207/
14:49:42 <john-davidge> Yep, it got some good attention over the weekend and is now in need of a new patch
14:49:50 <carl_baldwin> Maybe I'll try to spin it to address feedback.
14:50:04 <carl_baldwin> I know haleyb has his hands full.
14:50:34 <john-davidge> Does he need someone else to pick it up? I'd be happy to
14:51:48 <carl_baldwin> Let me do a quick turn on the spec.  We might need your help with the implementation.
14:52:16 <john-davidge> Sure
14:52:21 <carl_baldwin> #topic Open Discussion
14:52:46 <haleyb> carl_baldwin: i need to update, will do today
14:53:15 <carl_baldwin> haleyb: cool, let me know.  We're glad to help.
14:53:34 <john-davidge> haleyb: Feel free to ping me if you'd like to offload it to someone :)
14:55:10 <haleyb> if i could only find the right template :)
14:55:21 <carl_baldwin> I think that's it.  Thanks guys.  I'll be offline to commute to work.
14:55:39 <carl_baldwin> #endmeeting