15:00:41 <carl_baldwin> #startmeeting neutron_l3
15:00:45 <openstack> Meeting started Thu Nov 19 15:00:41 2015 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:49 <openstack> The meeting name has been set to 'neutron_l3'
15:01:11 <neiljerram_bb> Hi
15:01:34 <carl_baldwin> #topic Announcements
15:02:12 <yamamoto_> hi
15:02:15 <carl_baldwin> Mitaka-1 is fast approaching.
15:02:16 <carl_baldwin> #link https://wiki.openstack.org/wiki/Mitaka_Release_Schedule
15:02:24 <carl_baldwin> I can't believe it.
15:02:47 * regXboi hums Bob Dylan
15:03:06 <carl_baldwin> Any other announcements?
15:03:18 <juno_zhu> what features target to release in M-1?
15:04:42 <regXboi> #link https://launchpad.net/neutron/+milestone/mitaka-1 current milestone
15:04:56 <mlavalle> juno_zhu: ^^^^^
15:04:56 <carl_baldwin> juno_zhu: https://launchpad.net/neutron/+milestone/mitaka-1
15:05:05 <carl_baldwin> regXboi: jinx
15:05:06 <juno_zhu> thanks
15:05:16 <regXboi> carl_baldwin: no worries
15:05:46 <carl_baldwin> I'd also like to point out armax 's post to the ML yesterday
15:05:53 <carl_baldwin> #link http://lists.openstack.org/pipermail/openstack-dev/2015-November/079816.html
15:06:39 <carl_baldwin> A reminder to us all to prioritize our reviews to align focus.
15:07:32 <carl_baldwin> I you have blueprints targeted at Mitaka, be sure to be in contact with the approver for prompt review of your patches.
15:08:04 <carl_baldwin> If you are an approver, be sure to stay on top of those reviews especially.
15:08:31 <carl_baldwin> Anything else?
15:09:04 * regXboi shakes head
15:09:09 <carl_baldwin> #topic Bugs
15:09:14 <carl_baldwin> mlavalle: ping
15:09:18 <mlavalle> hi
15:09:31 <carl_baldwin> How is our bug situation looking?
15:09:48 <mlavalle> This past week we kept the same two bugs we had last meeting and added one, Although they are all getting reviews and more revisions
15:09:59 <mlavalle> first up is https://bugs.launchpad.net/neutron/+bug/1478100
15:09:59 <openstack> Launchpad bug 1478100 in neutron "DHCP agent scheduler can schedule dnsmasq to an agent without reachability to the network its supposed to serve" [High,In progress] - Assigned to Cedric Brandily (cbrandily)
15:10:40 <mlavalle> zzelle and amuller are jointly working on a fix https://review.openstack.org/#/c/205631/
15:11:02 <mlavalle> zzelle clasified it as a rfe
15:11:38 <carl_baldwin> This is related to the routed networks work too, actually.
15:11:57 <mlavalle> I don't think either of them are on-line right now
15:12:08 <mlavalle> any other comments?
15:12:43 <mlavalle> ok, moving on
15:12:54 <mlavalle> next up is https://bugs.launchpad.net/neutron/+bug/1494351
15:12:54 <openstack> Launchpad bug 1494351 in neutron "Observed StaleDataError in gate-neutron-dsvm-api tests if reference IPAM driver is used" [High,In progress] - Assigned to Pavel Bondar (pasha117)
15:13:15 <mlavalle> this is the bug that pavel_bondar has been fixing over the past few weeks
15:13:47 <mlavalle> https://review.openstack.org/#/c/223123/
15:14:15 <mlavalle> is got a fresh revision early today and it is awaiting reviews
15:14:42 <mlavalle> pavel_bondar: any comments?
15:15:43 <mlavalle> ok, next up is https://bugs.launchpad.net/neutron/+bug/1514810
15:15:43 <openstack> Launchpad bug 1514810 in neutron "Turning on 'enable_dhcp' on subnet update cause request failure for pluggable IPAM" [High,In progress] - Assigned to Gary Kotton (garyk)
15:16:09 <mlavalle> this one was reclassified by garyk to high a couple of days ago
15:16:10 <carl_baldwin> mlavalle: I'll review 223123 today.  It'll be on my priority list.
15:16:42 <mlavalle> and he proposed fix right away https://review.openstack.org/#/c/247019/
15:17:12 <mlavalle> awaiting reviews, it already got a few +1's
15:17:56 <mlavalle> any comments?
15:17:56 <carl_baldwin> Good to know.  I hadn't seen the proposed fix yet.
15:18:07 <johnbelamaric> it pulls out the validation
15:18:33 <johnbelamaric> because the validation was meaningless anyway - another thread could allocate the port while the validation was running
15:18:36 <carl_baldwin> I'm glad to see that John and Pavel have already looked at it from the IPAM perspective
15:19:13 <johnbelamaric> but that is going to re-open whatever bug the validation was put in to fix (evne though the validation didn't really fix it in a real world environment with multiple threads)
15:20:25 <mlavalle> and those are all the bugs I have this week. Any other bugs I might be missing?
15:20:35 <neiljerram_bb> Was the validation put in for a bug? It might have been just a mistaken assumption.
15:21:45 <johnbelamaric> I think it was, to fix https://launchpad.net/bugs/1443798
15:21:45 <openstack> Launchpad bug 1443798 in neutron juno "Restrict netmask of CIDR to avoid DHCP resync" [Undecided,Incomplete]
15:21:45 <carl_baldwin> johnbelamaric: Interesting.  I'll keep that in mind when I take a look at the patch.
15:22:24 <johnbelamaric> Look in garyk's commit message he reference the commit where the validation was put in
15:22:54 <carl_baldwin> johnbelamaric: ack
15:23:06 <carl_baldwin> mlavalle: I don't have any other bugs.  Anyone else?
15:23:11 <johnbelamaric> carl_baldwin: ok, as I say that I realize i didn't check if the transaction was open and would prevent the other thread allocating a port during the validation
15:24:51 <carl_baldwin> #topic Routed network segments
15:25:38 <carl_baldwin> This is actually progressing.  I am mostly finished with a spec update which I believe will make several people much happier.
15:26:23 <carl_baldwin> My updates will throw out the host <-> network mapping addition and describe how we can use bridge_mappings initially to fill the same need.
15:27:30 <carl_baldwin> They will also incorporate suggestions from neiljerram_bb to work in networking-calico use case to the model.
15:27:48 <neiljerram_bb> I liked your blog post on this, by the way.
15:28:12 <carl_baldwin> neiljerram_bb: Thanks
15:28:58 <neiljerram_bb> And then we'll all go and code it up!
15:29:16 <carl_baldwin> The updates will also include significant improves based on some feedback from johnthetubaguy from a Nova perspective.
15:30:06 * johnthetubaguy lurks in case of any questions
15:30:19 <carl_baldwin> It turns out the section on Nova / Neutron interaction wasn't in sync with Nova's plans.  I am reworking it and will try to iterate with John and Nova a bit more.
15:30:41 <carl_baldwin> johnthetubaguy: No questions yet.  Thanks for the feedback on the spec.
15:31:23 <johnthetubaguy> no problems, sorry we haven't been great at communicating our plans around the API and neutron integration, will try write up something so its more widely known
15:31:28 <carl_baldwin> #link https://review.openstack.org/#/c/225384/9/specs/mitaka/booting-vms-to-routed-networks.rst
15:31:45 <carl_baldwin> ^ That's the direct link to the spec btw in case anyone is curious.
15:32:02 <neiljerram_bb> Is there good doc anywhere of the flows between Nova and Neutron?
15:32:14 <carl_baldwin> johnthetubaguy: Great, I'll watch for that.
15:32:50 <carl_baldwin> neiljerram_bb: Actually, that is a good question to start with.  johnthetubaguy ^
15:33:27 <asadoughi> i don't know if it counts as docs, but this is *the* file https://github.com/openstack/nova/blob/master/nova/network/neutronv2/api.py
15:33:48 <carl_baldwin> asadoughi: That is where I've started to look, actually.
15:34:20 <neiljerram_bb> So that file would tell us all the ways that Nova *can* call Neutron?
15:35:10 <neiljerram_bb> I wonder if Neutron sometimes calls Nova too.
15:35:24 <neiljerram_bb> But anyway perhaps we should move on...
15:35:48 <asadoughi> there might be other places, too, would have to grep for that.
15:36:23 <asadoughi> and yes, i think there's neutron -> nova for port notifications
15:36:26 <carl_baldwin> neiljerram_bb: It can, I don't remember where the code lives off-hand.
15:36:52 <asadoughi> https://github.com/openstack/neutron/blob/master/neutron/notifiers/nova.py
15:37:20 <carl_baldwin> The other thing that I wanted to mention is I have some initial code out for review.  Thanks to reviewers for the reviews I've gotten.
15:37:23 <carl_baldwin> #link https://review.openstack.org/#/c/242270/
15:37:38 <neiljerram_bb> I started a related ML thread on 'REST API', motivated by trying to see these flows. So if anyone has insight, please chip in there.
15:37:57 <carl_baldwin> neiljerram_bb: Thanks.
15:38:33 <carl_baldwin> I wonder if anyone could help me with this patch set.  My own progress is slow on it because of a number of other responsibilities.
15:38:35 <asadoughi> i'd be happy to take a look. i've dealt with neutron <-> nova a lot here.
15:38:48 <tidwellr> carl_baldwin: I can help
15:39:07 <neiljerram_bb> I expect to have proper time for this, from Monday.
15:39:11 <carl_baldwin> tidwellr: Thanks.  That will be very much appreciated.
15:39:13 <tidwellr> my last attempt at a patch set was a fail, but it helped make progress
15:39:31 <tidwellr> :)
15:39:33 <carl_baldwin> tidwellr: I wouldn't call that a failure.
15:40:17 <carl_baldwin> tidwellr: Just had to change directions a bit.  It gave you valuable insight in to the bridge mappings discussion.
15:40:42 <tidwellr> yes it did, I like that direction a lot btw
15:41:07 <carl_baldwin> asadoughi: It would be great to get some insight in thin neutron <-> nova area.  I'd appreciate it.
15:41:51 <asadoughi> carl_baldwin: for sure. i'll be on it.
15:42:21 <carl_baldwin> asadoughi: Maybe when you've had a chance to look at it, we can talk again in the neutron channel.
15:42:30 <asadoughi> sounds good.
15:43:37 * mlavalle knows asadoughi really knows interaction between neutron and nova
15:43:42 <carl_baldwin> I'm going to shift my thinking toward attaching ports to an IpNetwork without a Network and without an IP address.  We need to discover what breaks down or goes wrong.
15:44:26 <neiljerram_bb> Without an IP address? You mean so that the IP address can be allocated later?
15:44:39 <carl_baldwin> neiljerram_bb: Yes.
15:44:47 <asadoughi> i like the idea if you're implying asynchronousity
15:44:54 <neiljerram_bb> Cool, sounds good
15:45:54 <carl_baldwin> I also had another thought for you.  I'll probably catch you a bit later so that we can move on.  We only have 15 minutes left!
15:46:00 <carl_baldwin> neiljerram_bb: ^
15:46:05 <neiljerram_bb> Sure
15:46:11 <carl_baldwin> That is all I have now.
15:46:26 <carl_baldwin> #topic BGP dynamic routing
15:46:43 <carl_baldwin> tidwellr: ping
15:47:50 <tidwellr> 1 last patch set to move some changes through https://review.openstack.org/#/c/216994/, then I'm going back through and adding address scope awareness
15:48:44 <tidwellr> based on some feedback from the summit, I'm not writing a service plugin and I need that change to ripple through the review chain
15:48:50 <tidwellr> not much else to report
15:49:48 <carl_baldwin> tidwellr: Could you use reviews?
15:50:14 <tidwellr> yes, those are helpful so we don't go off the rails
15:50:40 <carl_baldwin> tidwellr: I'll take a pass through today or tomorrow.
15:51:12 <carl_baldwin> #action carl_baldwin will review BGP patches
15:51:22 <carl_baldwin> #topic DNS
15:51:28 <carl_baldwin> mlavalle: ping
15:51:30 <mlavalle> hi again
15:51:43 <mlavalle> I pushed next revision to https://review.openstack.org/#/c/212213/
15:51:52 <carl_baldwin> sounds like the Nova BP is close
15:52:14 <mlavalle> I complted coding all the functionality for public ports on provider networks
15:52:52 <mlavalle> I need to debug it but am confident that we are close to the end
15:52:58 <carl_baldwin> mlavalle: cool
15:53:09 <mlavalle> garyk reviewed it right away
15:53:40 <mlavalle> he doesn't like the "shim" extension concept that I've been using. So I think I'll have to re-work it
15:54:35 <mlavalle> and yes, on the Nova side, johnthetubaguy indicated in https://review.openstack.org/#/c/90150/ that he feels it is close to be approved
15:54:52 <carl_baldwin> mlavalle: Need any help with the extension part?
15:55:28 <mlavalle> I incorporated his feedbak and carl_baldwin's, so I am awaiting reviews now
15:55:58 <mlavalle> carl_baldwin: as far as the extension, let me struggle with it over the next couple of days. I'll yell if I need help
15:56:37 <mlavalle> and that's my update for today
15:56:41 <carl_baldwin> mlavalle: good
15:56:52 <carl_baldwin> #topic address scopes
15:57:32 <carl_baldwin> The server side patches have been reviewed by salv-orl_ (the BP approver).  I need cores to look.
15:57:42 <carl_baldwin> haleyb: ^
15:58:06 * haleyb wakes up
15:58:25 <carl_baldwin> I'll ping ZZelle too or see who else I can round up
15:59:05 <juno_zhu> one router connects 2 different address scopes is allowed, right?
15:59:42 <carl_baldwin> juno_zhu: Right.  It is allowed to connect but traffic won't route freely between the two.
15:59:52 <salv-orl_> carl_baldwin: I think I've also reviewed one of the agent patches. That looks ok to me (I won't even argue about why we need a class to wrap directory creation ;) )
15:59:55 <haleyb> https://review.openstack.org/#/c/189948/ must be the review
16:00:19 * carl_baldwin looks at clock.
16:00:21 <carl_baldwin> #endmeeting