14:01:09 <carl_baldwin> #startmeeting neutron_routed_networks
14:01:10 <openstack> Meeting started Tue Mar  8 14:01:09 2016 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:14 <openstack> The meeting name has been set to 'neutron_routed_networks'
14:01:30 <carl_baldwin> mlavalle: reedip__: xiaohhui : Hi, and welcome.
14:01:41 <reedip__> hi carl_baldwin o/
14:01:53 <carl_baldwin> #topic Announcements
14:01:57 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-Routed-Networks#Agenda
14:02:14 <carl_baldwin> The blueprint for routed networks has been approved for Newton!
14:02:29 <mlavalle> ++
14:02:33 <carl_baldwin> #link https://blueprints.launchpad.net/neutron/+spec/routed-networks
14:02:48 <reedip__> gr8
14:02:54 <xiaohhui> That is great
14:02:57 <carl_baldwin> The RFE actually had been approved by the drivers for some time
14:03:00 <carl_baldwin> #link https://bugs.launchpad.net/neutron/+bug/1458890
14:03:00 <openstack> Launchpad bug 1458890 in neutron "Add segment support to Neutron" [Wishlist,Triaged] - Assigned to Carl Baldwin (carl-baldwin)
14:03:16 <carl_baldwin> The spec is looking pretty good too.
14:03:41 <carl_baldwin> #link https://review.openstack.org/225384
14:04:03 <carl_baldwin> It might be missing some detail but I think the overall picture is good.
14:04:14 <carl_baldwin> It is pretty cool to see the pathway forward opening up.
14:04:41 <xiaohhui> I went through the spec, and I agree with that
14:04:44 <carl_baldwin> Any other announcements?
14:04:56 <carl_baldwin> xiaohhui: Thanks, I saw your review and it is very much appreciated.
14:05:27 <carl_baldwin> #topic Progress
14:05:42 <reedip__> I havent seen the recent copy, will see it tomorrow morning ( its 11pm here)
14:05:58 <carl_baldwin> All the code should be posted to the gerrit topic
14:05:59 <carl_baldwin> #link https://review.openstack.org/#/q/status:open+topic:bp/routed-networks
14:06:02 <carl_baldwin> reedip__: Thanks.
14:07:02 <carl_baldwin> I added one more patch since last week but it is only a start.  It begins to link the subnets and the segments.
14:07:40 <carl_baldwin> However, I needed to go back and close a few things on the segment API / CRUD patch.
14:08:12 <carl_baldwin> First, I rebased it on top of a patch that moves the segments table out of the ML2 port of the code.  That was started by gongysh .
14:08:34 <carl_baldwin> I got that done and begin to reconcile the _db.py files from the two patches.
14:09:05 <carl_baldwin> So, that's where I am.  I didn't get a chance to work on it at all yesterday and may not again today.  :(
14:09:07 <vikram_> hi
14:09:16 <vikram_> sorry to be late
14:09:20 <carl_baldwin> vikram_: welcome
14:09:33 <xiaohhui> Carl, do you need a quick patch at neutronclient for the segment CRUD? I think I can forward the request to Richard
14:09:59 <vikram_> I can also offer help ;) if needed
14:10:10 <mlavalle> carl_baldwin: as soon as I finish with the unit tests we talked about yesterday I'll rebase my patchset
14:10:38 <reedip__> xiaohhui :Actually I would liKe to do the NeutronClient code, if its ok :)
14:10:45 <carl_baldwin> xiaohhui: That is a good question.  Someone at the Neutron mid-cycle offered to do that along with the openstack client but I haven't heard anything.
14:11:20 <carl_baldwin> xiaohhui: Right, it was Richard Theis.  Do you know his nick?
14:11:27 <reedip__> carl_baldwin: Richard did that , as per the notes of the mid-cycle circulated by Aramando
14:11:44 <xiaohhui> I can't get it now, but we can communicate inside IBM.
14:12:32 <carl_baldwin> xiaohhui: Thanks, it will be nice to have a client.  Currently, I'm using a generic REST client which works out nicely in the beginning but will be cumbersome in the long run.
14:13:04 <mlavalle> carl_baldwin: rtheis
14:13:31 <xiaohhui> OK, I see. I will send him a mail about it.
14:13:31 <carl_baldwin> reedip__: Maybe you can stay in communication with xiaohhui and rtheis and pick it up if there is an issue.
14:13:34 <carl_baldwin> mlavalle: Thanks.
14:13:37 <vikram_> carl_baldwin: +1.. I also believe let's have the server functionality up first.. then we can have client
14:13:55 <reedip__> carl_baldwin, xiaohhui: sure , thanks
14:13:59 <carl_baldwin> mlavalle: thanks
14:14:32 <carl_baldwin> mlavalle: Any progress report on https://review.openstack.org/285548 ?
14:15:19 <mlavalle> carl_baldwin: I'll rebase it later today as soon as I finish with the unit tests for DNS
14:15:30 <carl_baldwin> mlavalle: I had an action item to work on the host / subnet interface that I haven't followed through on yet.
14:15:45 <mlavalle> carl_baldwin: as you know, I had to catch up with those unit tests
14:15:54 <carl_baldwin> mlavalle: Thanks.  BTW, are those ready for review yet?
14:16:42 <mlavalle> carl_baldwin: they are, except for a couple of test cases for janzian changes
14:16:53 <carl_baldwin> #action carl_baldwin will follow through on host / subnet mapping idea (as opposed to lower level host / segment mapping only)
14:16:59 <mlavalle> carl_baldwin: but he needs to rebase one of his patchsets
14:17:08 <carl_baldwin> mlavalle: ok
14:17:30 <carl_baldwin> #action xiaohhui will follow up with rtheis on client changes
14:18:17 <reedip__> kindly keep me in the loop as well , would appreciate it
14:18:47 <carl_baldwin> xiaohhui: This host / segment mapping is relevant to OVN.  It will be required of the OVN plugin to provide it through the same interface we add for the ML2 plugin.
14:19:27 <mlavalle> carl_baldwin: also this week I am starting lookin at deferred IP allocation
14:19:56 <carl_baldwin> mlavalle: Good segue
14:20:08 <xiaohhui> I will investigate it in OVN
14:20:27 <carl_baldwin> xiaohhui: The interface is not well defined yet, maybe you could help us define that.
14:20:32 <rtheis> hi
14:20:45 <carl_baldwin> rtheis: Your ears burning?
14:20:54 <rtheis> they are :)
14:21:08 <carl_baldwin> rtheis: We were just talking about the client changes for the segment API.
14:21:18 <mlavalle> rtheis: Minessotta cold weather relieves that, doeesn't it?
14:21:28 <rtheis> it does
14:21:33 <carl_baldwin> .... which is still a bit of a WIP here:  https://review.openstack.org/284440
14:21:55 <rtheis> carl_baldwin: Is that what I need to get started with a CLI patch
14:23:06 <carl_baldwin> rtheis: yes
14:23:17 <rtheis> We need to lay some ground work to get neutronclient to function as an OSC plugin...
14:23:47 <rtheis> I'll need to look into that so I may try putting the CLI directly in OSC to start
14:24:12 <rtheis> I'll start working on this today
14:24:21 <carl_baldwin> rtheis: thanks
14:24:42 <carl_baldwin> #action rtheis to work on openstack CLI
14:25:01 <reedip__> rtheis, if OSC takes some time, then can we start implementing this on NeutronClient , in the meanwhile ?
14:25:13 <reedip__> if it is required
14:25:51 <carl_baldwin> reedip__: I think that sounds reasonable
14:26:14 <rtheis> reedip__: yes, that's fine
14:26:15 <reedip__> then I can work on that
14:26:25 <rtheis> reedip__: thank you
14:26:45 <carl_baldwin> xiaohhui: Do you think that working on the plugin interface for mapping is something you could tackle?
14:26:49 <reedip__> rtheis, carl_baldwin: great :)
14:27:00 <xiaohhui> carl_baldwin: I think I can do that.
14:27:11 <carl_baldwin> ... maybe working from https://review.openstack.org/285548 as a start
14:27:29 <carl_baldwin> #action xiaohhui will work on plugin interface for host mapping
14:27:39 <carl_baldwin> mlavalle: back to your seque ...
14:28:22 <carl_baldwin> I think deferred IP allocation is a very good thing to start playing with early.
14:28:29 <mlavalle> agree
14:28:55 <carl_baldwin> mlavalle: What is your plan to get that started?
14:29:56 <mlavalle> carl_baldwin: don't have a plan yet. I'll start going back to the spec and then lookin at the relevant parts in Nova
14:31:30 <carl_baldwin> mlavalle: I think what we really need to do is get the subnet / segment association patch to the point where IPAM actually defers allocation and see what that does with the various Nova boot use cases in the Nova spec.
14:32:30 <carl_baldwin> mlavalle: But, maybe you don't need all of that.
14:33:06 <mlavalle> carl_baldwin: I'll follow that path... But I will start lookin on the Nova side any way
14:33:10 <carl_baldwin> mlavalle: You can probably fake it for now by just removing the IPAM allocation until a host is know.
14:33:13 <carl_baldwin> *known
14:34:03 <carl_baldwin> mlavalle: I don't want to make this too dependent on other work because I think it would be good for you to start down the Nova path sooner than later.
14:34:05 <mlavalle> carl_baldwin: I think your point is very good. At the very least I'll keep an eye on the subnet / segment assoc patch
14:34:11 * carl_baldwin just thinking out loud.
14:34:37 <carl_baldwin> mlavalle: There are going to be some IPAM complications to work out with that patch which I can get started on.
14:34:38 <mlavalle> carl_baldwin: but yes, I think starting on Nova ASAP is paramount
14:35:17 <carl_baldwin> #action carl_baldwin will work on IPAM complications around subnet / segment association.  Will loop in johnbelamaric .
14:35:37 <carl_baldwin> #action mlavalle will work on deferred IP allocation as independently as possible
14:35:41 <carl_baldwin> mlavalle: That sound good?
14:36:00 <mlavalle> carl_baldwin: Yes
14:36:22 <carl_baldwin> mlavalle: cool
14:37:08 <carl_baldwin> My biggest concern still lies around following Nova's scheduler work as a moving target.  mlavalle that's why I think you should steer in that direction.
14:37:20 <carl_baldwin> ... also making sure we can get and keep Nova's attention on this.
14:38:02 <mlavalle> carl_baldwin: yeah, one of my team members at work is leading the scheduler team now. So I will have help from him
14:38:42 <mlavalle> carl_baldwin: I already mentioned this to him (edleafe)
14:39:21 <carl_baldwin> mlavalle: Great.  I actually watched their meeting yesterday morning and was about to send a reminder about the spec but they ended the meeting a bit eraly.
14:39:23 <carl_baldwin> *early
14:39:33 <edleafe> carl_baldwin: you have my attention :)
14:39:46 <carl_baldwin> edleafe: hi, welcome
14:39:58 <mlavalle> edleafe: thanks :-)
14:40:20 <edleafe> always happy to help make things run smoothly
14:40:56 <carl_baldwin> edleafe: I don't think we met at the Nova mid-cycle in Bristol.  Were you there?
14:41:21 <edleafe> carl_baldwin: no, couldn't get travel approved until it was way too late
14:41:31 <edleafe> carl_baldwin: will you be in Austin?
14:41:38 <carl_baldwin> edleafe: I will be in Austin.
14:41:45 <edleafe> kewl
14:41:53 <carl_baldwin> edleafe: Look forward to meeting you.
14:42:10 <carl_baldwin> edleafe: You missed some great discussions about this subject at the mid-cycle.  We'll need to catch you up.
14:42:16 <edleafe> likewise
14:42:36 <edleafe> yeah, I am still amazed at how disconnected I felt after the midcycle
14:43:02 <edleafe> I'll have to fight harder to get the powers that be to approve travel. It's too important
14:43:18 <carl_baldwin> mlavalle: edleafe: let me know if I can help catching you up.
14:43:32 <edleafe> anyway, I don't want to derail the meeting. mlavalle can fill me in
14:43:33 <carl_baldwin> edleafe: A good start will be the Nova spec I wrote.
14:43:46 <carl_baldwin> #link https://review.openstack.org/263898
14:44:09 <mlavalle> carl_baldwin: edleafe is a San Antonian like me. So we could meet face to face to talk about it
14:44:10 <edleafe> carl_baldwin: added to my list
14:44:22 <carl_baldwin> edleafe: No worries.  This fit well in to the meeting.
14:45:02 <carl_baldwin> #topic OVN and Routed Networks
14:45:27 <carl_baldwin> I'm not sure if russellb is lurking ...
14:46:47 <carl_baldwin> xiaohhui: Anything else on this topic that we haven't discussed already?
14:46:48 <russellb> hi
14:46:49 <russellb> sorry
14:47:02 <russellb> on phone, thought this started in 15 minutes, was planning to attend
14:47:10 <russellb> off phone!
14:47:29 <russellb> i think we covered it pretty well in person
14:47:34 <russellb> architecturally i think we can support it
14:47:35 <carl_baldwin> russellb: Sorry, it changes time every week.  Just to make things hard.  ;)
14:47:41 <russellb> comes down to making sure someone keeps up and does the work
14:47:47 <russellb> and i think mestery was able to idenitfy some people to help with that
14:47:57 <russellb> goal is to support it with OVN out of the gate
14:48:03 <mestery> Yes, they should be on it now!
14:48:04 <russellb> seems doable
14:48:30 <carl_baldwin> russellb: mestery:  I think we have a good start this week with xiaohhui
14:48:36 <russellb> awesome
14:48:43 <mestery> nice
14:48:46 <xiaohhui> :)
14:48:50 <russellb> xiaohhui: thank you!
14:48:59 <carl_baldwin> Anything else?
14:49:13 <russellb> xiaohhui: reach out if you hit anything needed from OVN
14:49:38 <xiaohhui> russellb: I will, thanks
14:50:02 <carl_baldwin> Thanks!
14:50:08 <carl_baldwin> #topic Router external gateways on routed networks
14:50:22 <carl_baldwin> This topic is a bit more future looking ...
14:50:51 <carl_baldwin> #action carl_baldwin will start a spec on this
14:51:07 <carl_baldwin> That's really all I had to say about this for now.
14:51:13 <carl_baldwin> #topic Open Discussion
14:51:16 <carl_baldwin> Anything else?
14:51:35 <mlavalle> excited to see this getting off the ground! yaay
14:52:54 <carl_baldwin> So, I'm on vacation next week.  I'll be in Washington DC and the meeting was to be at 1500.
14:54:05 <mlavalle> carl_baldwin: we move to dst this week, don't we?
14:54:20 <carl_baldwin> mlavalle: Yep.
14:54:30 <carl_baldwin> Should we hold the meeting next week?
14:54:37 <reedip__> oh so what would be the new time?
14:55:09 <mlavalle> reedip__: it doesn't change for you. This is US moving to day saving time
14:55:16 <carl_baldwin> The meeting times are all in UTC.
14:55:21 <carl_baldwin> ... thank heavens
14:56:08 <reedip__> ohyrah
14:56:14 <reedip__> i forgot
14:56:30 <carl_baldwin> mlavalle: Will you chair the meeting next week?  I'll watch from my phone or catch the logs.  It will be at 11:00 am in DC.
14:56:41 <mlavalle> carl_baldwin: will do
14:57:07 <carl_baldwin> If you're in a really late time zone, feel free to go to bed early and email progress or something.
14:57:07 <mlavalle> carl_baldwin: let's synch up Friday afternoon
14:57:19 <carl_baldwin> But, it will be good to see what progress has been made.
14:57:22 <carl_baldwin> mlavalle: ack.
14:58:03 <carl_baldwin> That's all folks!
14:58:17 <carl_baldwin> #endmeeting