15:00:02 <carl_baldwin> #startmeeting neutron_l3
15:00:03 <openstack> Meeting started Thu Jul  9 15:00:02 2015 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:07 <openstack> The meeting name has been set to 'neutron_l3'
15:00:11 <carl_baldwin> #topic Announcements
15:00:16 <pavel_bondar> hi
15:00:18 <johnbelamaric> hi
15:00:18 <haleyb> hi
15:00:21 <dlundquist> hi
15:00:37 <tidwellr> hi
15:00:38 <carl_baldwin> Any announcements?
15:01:22 <carl_baldwin> The name for M was *almost* chosen, but not.
15:01:36 <regXboi> but not?
15:01:39 <mlavalle> yeah, it was rescinded yesterday
15:01:44 <neiljerram> hi
15:01:49 <regXboi> what was the issue?
15:02:17 <carl_baldwin> I’m trying to find the ML link.
15:02:17 <john-davidge> unintended negative curtural meaning, it seems
15:02:26 <john-davidge> *cultural
15:02:34 <neiljerram> It was the name of a Japanese emperor who invaded Korea and other places
15:02:49 <regXboi> oy
15:02:50 <neiljerram> (Hoping I'm guessing the correct conversation here...)
15:02:59 <carl_baldwin> Any other announcements?
15:03:16 <carl_baldwin> #topic Bugs
15:03:34 <mlavalle> carl_baldwin: this is the M email http://lists.openstack.org/pipermail/openstack-dev/2015-July/068935.html
15:03:49 <carl_baldwin> mlavalle: Thanks.
15:04:03 <carl_baldwin> The only bug on the list right now in bug 1404743
15:04:03 <openstack> bug 1404743 in neutron "sporadic test failures due to VMs not getting a DHCP lease" [High,Confirmed] https://launchpad.net/bugs/1404743
15:04:14 <vikram> Hi All
15:04:20 <vikram> Sorry to be late:-)
15:05:02 <carl_baldwin> I found that the problem still occurs but with less frequency that 6 months ago.  I did not have time to look further.  This is the type of thing which could have more than one cause over time.
15:05:21 <carl_baldwin> I did not look further in to the problem.
15:05:26 <carl_baldwin> Any other bugs to bring up?
15:05:40 <amuller> carl_baldwin: https://bugs.launchpad.net/neutron/+bug/1465434
15:05:40 <openstack> Launchpad bug 1465434 in neutron "DVR issues with supporting multiple subnets per network on DVR routers - PortbindingContext does not have the status." [Critical,In progress] - Assigned to Swaminathan Vasudevan (swaminathan-vasudevan)
15:06:01 <amuller> carl_baldwin: if I'm not mistaken that's the bug that's had the DVR job failure rate at 99% the last month and change
15:06:15 <amuller> I feel like the patch is not getting as much attention as it deserves
15:06:38 <john-davidge> this might be a good place to bring up https://bugs.launchpad.net/neutron/+bug/1471316
15:06:38 <openstack> Launchpad bug 1471316 in neutron "_get_subnetpool_id does not return None when a cidr is specified and a subnetpool_id isn't." [High,In progress] - Assigned to John Davidge (john-davidge)
15:06:58 <amuller> carl_baldwin: But I see that you and Brian reviewed it recently so it's hard to complain :)
15:07:24 <carl_baldwin> amuller: Thanks.  You think it could use more attention?  From reviewers or the author?
15:07:49 <haleyb> amuller: yes, i've been reviewing it, and it still needs some more work
15:07:55 <amuller> ok
15:08:04 <amuller> a lot of people are looking at the DVR job failure rate
15:08:19 <amuller> this is almost a marketing thing at this point
15:08:49 <mlavalle> carl_baldwin: I'll add it to our wiki page, so it is visible there
15:09:36 <carl_baldwin> mlavalle: Thanks.
15:09:48 * carl_baldwin will step out for a moment.  mlavalle to continue.
15:09:51 <carl_baldwin> #chair mlavalle
15:09:52 <openstack> Current chairs: carl_baldwin mlavalle
15:10:26 <mlavalle> amuller: I will add that bug to our wiki page and I'll track it continiusly
15:10:33 <amuller> mlavalle: Thank you
15:10:50 * mlavalle not spellin very well long words
15:11:07 <mlavalle> john-davidge: what about that bug
15:11:49 <john-davidge> mlavalle: The consensus so far seems to be that it’s not a bug, but the original spec disagrees
15:12:08 <john-davidge> mlavalle: see my most recent comment
15:12:30 * mlavalle looking
15:12:46 <john-davidge> We’re currently treating the default pool differently to any other pool, and i think that’s confusing
15:14:08 <john-davidge> I think that the default pool should be used in situations where the user is clearly indicating that they wish to use a pool, but they haven’t specified which one. Rather than overriding the default subnet-create behaviour when defined.
15:15:03 <mlavalle> john-davidge: does that imply code change?
15:16:04 <john-davidge> mlavalle: I beleive a code change is nessessary to go back to what’s defined in the spec, yes.
15:16:14 <john-davidge> I’ve proposed a fix https://review.openstack.org/#/c/198437/
15:16:43 <mlavalle> john-davidge: ahhhh, ok, let's make it sure you get reviews and try to get this to consesus
15:16:45 * carl_baldwin back
15:16:56 <john-davidge> mlavalle: Thanks :)
15:17:21 <mlavalle> carl_baldwin: all your now
15:17:23 <tidwellr> john-davidge: I haven't been following this closely, the None (or null) subnet pool behaves a little differently than a default pool. I'll take a look at the review. I think you'll have to default IPv4 and IPv6 to different pools if you the behavior you're describing
15:18:05 * tidwellr can't type a proper sentence
15:18:15 <john-davidge> tidwellr: There default pools for IPv4 and IPv6 are defined by separate config options
15:18:33 <tidwellr> ok, cool. I'll take a look at the review
15:18:44 <john-davidge> tidwellr: great, thanks
15:18:54 <carl_baldwin> john-davidge: I’ll have another look at your comment.
15:19:25 <john-davidge> carl_baldwin: Thanks Carl, hope I’m explaining my position well enough
15:19:41 <carl_baldwin> john-davidge: thanks
15:19:47 <carl_baldwin> Any other bugs to mention?
15:21:11 <carl_baldwin> #topic Routing Network Segments
15:21:27 <carl_baldwin> #link https://review.openstack.org/#/c/196812/
15:21:41 <carl_baldwin> Seems to have good consensus.  Need to get implementation started.
15:22:06 <regXboi> carl_baldwin: there is another spec out there that looks to have a bit of overlap
15:22:20 <carl_baldwin> regXboi: Which one is that?
15:22:30 * regXboi looking for it - routed networks
15:22:45 <neiljerram> carl_baldwin: I think that's mine.
15:23:25 <carl_baldwin> Actually, I was happy with the way this one and neiljerram ’s complement each other.
15:23:44 <neiljerram> But even if there does end up being overlap - e.g. at the API, or in implementation - I think it still works to proceed with 196812 first, and then do mine as a (possible) delta
15:24:05 <carl_baldwin> neiljerram: +1
15:24:19 <carl_baldwin> neiljerram: Do you have a link handy.  This is a good reminder for me to revisit it.
15:24:25 <regXboi> https://review.openstack.org/#/c/198439/
15:24:27 <neiljerram> carl_baldwin: Do you know what the work items are for 196812
15:24:29 * regXboi found it
15:24:55 <neiljerram> https://review.openstack.org/#/c/198439/
15:25:09 <neiljerram> D'oh, sorry for duplication!
15:25:14 <carl_baldwin> regXboi: neiljerram: Thanks.  Looks like there are more comments to read.
15:25:30 * carl_baldwin notices time is ticking away.  :)
15:25:38 <regXboi> carl_baldwin: there are, and I small bgp and AS lurking on the horizon :/
15:25:47 <regXboi> er *smell not small
15:25:54 <neiljerram> carl_baldwin: Yes.  Main thing I'd like to ask - later - is what correct process should be. But back to your spec now!
15:26:38 <carl_baldwin> neiljerram: Even I’m a bit confused about process.  Let’s talk about it when the meeting closes.
15:26:45 <carl_baldwin> #topic BGP dynamic routing
15:26:54 <neiljerram> carl_baldwin: OK, will do
15:27:08 <carl_baldwin> tidwellr: vikram_: hi
15:27:18 <vikram_> carl_baldwin: hi
15:27:22 <tidwellr> hi
15:27:31 <carl_baldwin> What’s the latest?
15:28:03 <tidwellr> I think I've a handle on how to setup callbacks on FIP associate/disassociate
15:28:54 <tidwellr> that will allow us to invoke RPC's on the dr_agent that will advertise/unadvertise the host routes
15:29:07 <carl_baldwin> tidwellr: Any progress on setting up a test environment?
15:29:37 <tidwellr> yes, I've got a couple quagga instances running and I actually have a real router that I think runs BGP
15:29:46 <tidwellr> finishing touches on it today
15:30:04 <carl_baldwin> tidwellr: cool.  Sounds like good progress.
15:30:05 <tidwellr> definitely something to document and share around
15:30:16 <tidwellr> taking notes as I go
15:31:10 <carl_baldwin> Anything else to discuss?
15:31:34 <tidwellr> vikram is working on the dr_agent in parallel I believe
15:31:40 <tidwellr> nothing for me
15:31:46 <Vikram> tidwellr:+1
15:31:59 <Vikram> will post a patch soon
15:32:00 <carl_baldwin> vikram: Any thing to discuss?
15:32:08 <carl_baldwin> vikram: Great.
15:32:19 <Vikram> nothing on this topic carl.. need to catch up :-)
15:32:25 <carl_baldwin> #topic IPAM
15:32:32 <carl_baldwin> pavel_bondar: johnbelamaric: hi
15:32:37 <johnbelamaric> carl_baldwin: hi
15:32:40 <pavel_bondar> hi
15:33:12 <pavel_bondar> updated today #link https://review.openstack.org/#/c/183099/
15:33:42 <johnbelamaric> I think you meant #link https://review.openstack.org/#/c/153236
15:33:53 <pavel_bondar> sorry, yes
15:34:06 <carl_baldwin> Just got it.  I had it up when the meeting started.
15:34:50 <pavel_bondar> and working on changes for the #link https://review.openstack.org/#/c/198777/
15:35:17 <johnbelamaric> alright looks like we may not hit PS 100
15:35:20 <johnbelamaric> :)
15:35:24 <pavel_bondar> :)
15:35:49 <carl_baldwin> pavel_bondar: I was just looking at your comments on 198777.
15:36:58 <carl_baldwin> But, it looks like things are still moving.  I’ll keep my eye on these patches.
15:37:18 <pavel_bondar> carl_baldwin: thanks
15:37:36 <carl_baldwin> #topic ML3 router plugin
15:37:42 <carl_baldwin> Any update here?
15:38:02 <yamahata> Hi. No much update
15:38:27 <carl_baldwin> yamahata: Thanks.
15:38:35 <yamahata> Now we are giving a consideration on implementation again based on use cases
15:38:47 <carl_baldwin> yamahata: ack
15:38:56 <carl_baldwin> #topic DNS
15:38:58 <mlavalle> carl_baldwin: earlier this week we merged this http://specs.openstack.org/openstack/neutron-specs/specs/liberty/external-dns-resolution.html
15:39:29 <carl_baldwin> mlavalle: It is great to see it merged.
15:39:45 <mlavalle> As far as coding for the internal side of it, I made very good progress over the past 3 days. I have all the circuit for create_port implmented from the api to the dhcp agent
15:40:30 <mlavalle> Yesterday morning I thought that the port_update was gling to be very easy, but it turned out to be a little trickey when you combine changes in ips and dns_name
15:40:40 <mlavalle> so I got bogged down a little bit yesterday
15:40:56 <mlavalle> I think I have a handle on it now. Expect to push code today or tomorrow
15:41:35 <mlavalle> and i am starting the implementation of the xternal part also over the next 2 days
15:41:58 <mlavalle> of course I keep sheperding the nova spec as well
15:42:10 <mlavalle> that is where I am
15:42:11 <carl_baldwin> mlavalle:  I am looking forward to reviewing it.
15:42:28 <carl_baldwin> mlavalle: Don’t let the nova spec slow you down.
15:42:42 <mlavalle> carl_baldwin: I won't, :-)
15:42:46 <carl_baldwin> :)
15:42:53 <carl_baldwin> mlavalle: Anything else?
15:42:58 <mlavalle> that's it for me
15:43:03 <carl_baldwin> #topic Address Scopes
15:43:06 <carl_baldwin> vikram: hi
15:43:45 <carl_baldwin> vikram: The patch for associated address scopes to subnet pools is coming along well, I think.
15:43:49 <carl_baldwin> Very good progress.
15:44:16 <carl_baldwin> #link https://review.openstack.org/197552
15:44:31 <carl_baldwin> I also began looking at the patch for API tests
15:44:37 <Vikram> thanks
15:44:40 <carl_baldwin> #link https://review.openstack.org/198850
15:45:12 <carl_baldwin> I haven’t looked much at the CLI patches yet but I see that others have.
15:46:33 <carl_baldwin> link:  https://review.openstack.org/#/q/topic:bp/address-scopes+status:open,n,z
15:46:36 <carl_baldwin> #link https://review.openstack.org/#/q/topic:bp/address-scopes+status:open,n,z
15:47:18 <carl_baldwin> I’m pretty close on the L3 agent implementation too.  The code is working locally but some of it needs work before posting.
15:47:47 <carl_baldwin> That’s all I have.
15:47:59 <Vikram_> carl_baldwin: let me know if need help
15:48:09 <carl_baldwin> Vikram_: Thanks.
15:48:13 <carl_baldwin> #topic IPv6
15:49:12 <carl_baldwin> haleyb: HenryG: john-davidge: others:  Anything more to discuss today?
15:49:30 <regXboi> carl_baldwin: I do :)
15:49:37 <sambetts> night dt
15:50:04 <john-davidge> latest Prefix Delegation DB patch has a couple of +1s right now https://review.openstack.org/#/c/158697/
15:50:05 <carl_baldwin> regXboi: ?
15:50:22 <regXboi> carl_baldwin: https://etherpad.openstack.org/p/neutron-instrumentation is getting fleshed out - looking for folks to review
15:50:45 <john-davidge> carl_baldwin: pavel_bondar: Thanks for the detailed reviews
15:51:07 <regXboi> carl_baldwin: hoping to spin a draft umbrella bp sometime next week
15:51:10 <carl_baldwin> john-davidge: I will look at it today.
15:51:52 <carl_baldwin> regXboi: ack
15:51:59 <carl_baldwin> #topic Open Discussion
15:52:19 <john-davidge> carl_baldwin: thanks
15:54:25 <haleyb> not neutron-specific, but we did just land a few changes to make devstack run over IPv6 (apis, etc), except for the OVS tunnels
15:55:01 <john-davidge> haleyb: cool!
15:55:34 <carl_baldwin> +1.  Great work!
15:58:30 <carl_baldwin> I guess that’s it.  Thanks all!
15:58:36 <mlavalle> bye
15:58:38 <neiljerram> bye!
15:58:41 <amuller> toodles
15:58:49 <regXboi> oom
15:58:54 <neiljerram> carl_baldwin: will catch you in about 30 mins, if that's OK
15:58:59 <carl_baldwin> #endmeeting