15:00:54 <carl_baldwin> #startmeeting neutron-l3
15:00:54 <openstack> Meeting started Thu Aug  6 15:00:54 2015 UTC and is due to finish in 60 minutes.  The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:55 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:55 <vikram_> hi
15:00:58 <openstack> The meeting name has been set to 'neutron_l3'
15:01:00 <pavel_bondar> hi
15:01:06 <numan> hi
15:01:10 <carl_baldwin> #topic Announcements
15:01:14 <carl_baldwin> #link https://wiki.openstack.org/wiki/Meetings/Neutron-L3-Subteam
15:01:21 <john-davidge> hi
15:01:27 <carl_baldwin> We’re in the final month of feature development.
15:02:17 <carl_baldwin> Liberty-3 will be at the end of this month.  Where has the time gone?
15:02:20 <carl_baldwin> Any other announcements?
15:02:45 <regXboi> only that we are trying to add multimode-dvr to the check queue as nonvoting
15:03:00 <regXboi> and once it is added, I'll be watching it to see if we can make it voting
15:03:08 <carl_baldwin> regXboi: Ah, yes.  Thanks.  Do you have a link to ther review?  I think I saw a +2 pass by.
15:03:24 <regXboi> #link https://review.openstack.org/#/c/209165/
15:03:36 <regXboi> and it has been W+1
15:03:55 <regXboi> so ... here we go
15:04:22 <carl_baldwin> regXboi: Excelent.  This should bring in some good data.
15:04:29 * regXboi crosses fingers :)
15:04:32 <carl_baldwin> Other announcements?
15:04:59 <carl_baldwin> #topic Bugs
15:05:02 <carl_baldwin> mlavalle: ping
15:05:18 * carl_baldwin thanks mlavalle for his efforts in keeping up on the meeting agenda and the bugs.
15:05:39 <mlavalle> carl_baldwin: so i followed up with the owner of https://bugs.launchpad.net/neutron/+bug/1388698
15:05:39 <openstack> Launchpad bug 1388698 in neutron "dhcp_agents_per_network does not work appropriately." [High,In progress] - Assigned to watanabe.isao (watanabe.isao)
15:06:10 <mlavalle> we didn't have any progreess report on that for a few konths
15:06:27 <mlavalle> it turns out we have a fix proposed for that and it seems it will merge soon
15:07:02 <mlavalle> also this coming week i intend to invest sometime researching https://bugs.launchpad.net/neutron/+bug/1404743
15:07:02 <openstack> Launchpad bug 1404743 in neutron "sporadic test failures due to VMs not getting a DHCP lease" [High,Confirmed]
15:07:13 <carl_baldwin> mlavalle: Thanks, I actually looked at that review recently.  Was waiting for yamamoto to revisit it.
15:07:52 <mlavalle> we need to dig deeper into that. i will ask some help.......
15:08:19 <mlavalle> yeah, if yamamoto can revisit https://bugs.launchpad.net/neutron/+bug/1388698
15:08:19 <openstack> Launchpad bug 1388698 in neutron "dhcp_agents_per_network does not work appropriately." [High,In progress] - Assigned to watanabe.isao (watanabe.isao)
15:08:32 <yamamoto> i will.
15:08:36 <mlavalle> that will help us to almost drive our bugs to 0
15:08:47 <carl_baldwin> Thanks, mlavalle !
15:08:57 <mlavalle> other than those 2 bugs, the rest are medium priority or lower
15:09:17 <regXboi> so... there is one new l3-dvr-backlog that got made high last week
15:09:36 <mlavalle> any other bugs folks wnat to bring up?
15:09:41 <regXboi> (me goes to look for the number
15:10:09 <carl_baldwin> regXboi: Thanks for the reminder.  I think this team needs to keep a better eye on that bug tag.
15:10:24 <regXboi> carl_badlwin: I'm sort of doing that
15:10:36 <carl_baldwin> At one point, l3-dvr-backlog was a subset of l3-dhcp-ipam but that is no longer true.
15:10:53 <carl_baldwin> (or is it l3-ipam-dhcp?)
15:11:11 <carl_baldwin> Anyway, I appreciate regXboi ’s help with keeping an eye on that.
15:11:18 <mlavalle> carl_baldwin, regXboi: iwans't watching the dvr one.... i'll start doing it from now on
15:11:21 <regXboi> https://review.openstack.org/#/c/209165/
15:11:26 <regXboi> sorry, that's the patch
15:11:33 <regXboi> #link https://bugs.launchpad.net/neutron/+bug/1405910
15:11:33 <openstack> Launchpad bug 1405910 in neutron "distributed router can not add routes" [High,In progress] - Assigned to Ryan Moats (rmoats)
15:11:59 <regXboi> there we go - the patch had been abandoned - I've picked it up and am trying to shepherd it through
15:12:25 <haleyb> and a silent failure, ouch, i will keep an eye as well
15:12:34 <carl_baldwin> regXboi: Many thanks.  We need a little reviewer attention on it.
15:13:02 <regXboi> carl_baldwin: ack - I need to add test cases to it
15:13:12 <regXboi> I'm still trying to get the rebase happy
15:13:18 <carl_baldwin> #link https://review.openstack.org/#/c/144205/
15:13:26 <carl_baldwin> ^ Reviewers
15:13:30 <carl_baldwin> Any other bugs?
15:13:47 <mlavalle> regXboi: thanks!
15:13:49 <regXboi> let's get the other ones on the table
15:14:14 <regXboi> #1365473
15:14:29 <regXboi> 1369721
15:14:37 <regXboi> 1470889
15:14:46 <regXboi> and now to make links
15:14:59 <regXboi> #link https://bugs.launchpad.net/neutron/+bug/1365473
15:14:59 <openstack> Launchpad bug 1365473 in neutron "Unable to create a router that's both HA and distributed" [High,In progress] - Assigned to Adolfo Duarte (adolfo-duarte)
15:15:05 <regXboi> #link https://bugs.launchpad.net/neutron/+bug/1369721
15:15:05 <openstack> Launchpad bug 1369721 in neutron "manually moving dvr-snat router fails" [High,In progress] - Assigned to Oleg Bondarev (obondarev)
15:15:14 <regXboi> #link https://bugs.launchpad.net/neutron/+bug/1470889
15:15:15 <openstack> Launchpad bug 1470889 in neutron "DVR: should not remove routers form down l3 agents on compute nodes" [High,In progress] - Assigned to Oleg Bondarev (obondarev)
15:15:35 <regXboi> iirc, those all have current patches...
15:15:56 <regXboi> that's all from me
15:15:59 <carl_baldwin> regXboi: Thanks.
15:16:18 <carl_baldwin> Here’s a link to the dvr backlog that we should add to our wiki page:
15:16:20 <carl_baldwin> #link https://bugs.launchpad.net/neutron/+bugs?field.tag=l3-dvr-backlog
15:16:39 <carl_baldwin> Yikes, looks like some triage is in order.
15:16:46 <regXboi> :(
15:16:57 <mlavalle> carl_baldwin: adding the tag now....
15:17:56 <regXboi> carl_baldwin: I'll try and spend some time checking to see what has patches and what doesn't and ...
15:18:21 <carl_baldwin> regXboi: Many many thanks.
15:19:12 <carl_baldwin> Moving on...
15:19:15 <carl_baldwin> #topic Routed network segments
15:19:57 <carl_baldwin> This topic sure seems to be heating up!  I love it but it is now clear that we’ve got some heavy lifting to do.
15:20:05 <carl_baldwin> #link https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg58876.html
15:20:52 <carl_baldwin> I’ve taken an action item to write a new proposal encorporating the feedback from this thread.
15:21:36 <carl_baldwin> I guess there isn’t much to discuss here without a more fleshed out proposal.
15:21:51 <carl_baldwin> #topic BGP dynamic routing
15:22:04 <carl_baldwin> vikram_: tidwellr:  ping
15:22:05 <tidwellr> hi
15:22:10 <vikram_> hi
15:23:03 <carl_baldwin> Actually, one more thing on routed networks.
15:23:04 <carl_baldwin> #undo
15:23:05 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x8294050>
15:23:25 <carl_baldwin> regXboi: Do you want to mention the operator’s meetup?
15:23:59 <regXboi> sure
15:24:01 <carl_baldwin> The details are buried in that thread that I linked.  I think it was going to be on Aug 18th.
15:24:21 <regXboi> carl_baldwin: that sounds correct - let me find the right link
15:24:27 <regXboi> I'll add it later, so we can move along
15:24:28 <tidwellr> I'm will be there, that's correct
15:24:41 <carl_baldwin> I will be at linuxcon with mestery and we plan to find a corner to sit it and join the discussion.
15:25:02 <mestery> ++
15:25:18 <carl_baldwin> My only concern is that it sounds like kevinbenton may not be able to join at the chosen time from China.
15:25:42 <carl_baldwin> kevinbenton ’s feedback has been valuable
15:25:43 <regXboi> #link http://lists.openstack.org/pipermail/openstack-dev/2015-August/071341.html
15:25:51 <regXboi> oops
15:26:04 <regXboi> that's the email with the proposed time on 8/18
15:26:24 <regXboi> carl_baldwin: that is unfortunate, but I think we'll have to muddle through
15:26:35 <carl_baldwin> Okay, let’s move on.
15:26:41 <carl_baldwin> #topic BGP dynamic routing
15:26:48 <carl_baldwin> vikram_: tidwellr : take two
15:27:01 <tidwellr> I've been going round and round with where to insert callback notifications, anyway we've settled just inserting them where we need them for now and dealing with notifications more broadly later
15:27:28 <vikram_> I am working on the unit tests
15:27:29 <tidwellr> I think I've cleared the way for us to wrap this up now
15:28:12 <carl_baldwin> tidwellr: vikram_:  Where do you need reviews?
15:28:44 <tidwellr> these reviews have gone a little quiet recently, mostly because I've been tied up and working through things in a sandbox
15:29:51 <tidwellr> I'm going to start re-building the review dependencies today or tomorrow, but start here https://review.openstack.org/#/c/201621/
15:30:08 <carl_baldwin> tidwellr: When they pick up, be sure to ping generously.
15:30:12 <vikram_> My patches has lot of failures.. Will correct them soon .. Probably need a weeks time to get it up for review
15:30:42 <tidwellr> we're so close, but we do need to things up for review
15:31:02 <carl_baldwin> tidwellr: vikram_: thanks.
15:31:12 <baoli> tidwellr: a quick question on BGP. What network are the BGP speakers and the BGP peers running on?
15:31:24 <tidwellr> they can run anywhere
15:31:53 <tidwellr> you can deploy the agent where you need it, as long as it can connect to rabbitmq and communicate with the server
15:32:07 <baoli> tidwellr, but the peer's address must be reachable freom where the dr agent running?
15:32:13 <tidwellr> yes
15:32:26 <baoli> tidwellr, ok, sounds like the mgmt net, then
15:32:36 <tidwellr> well, not necessarily
15:32:41 <tidwellr> depends on your environment
15:33:02 <tidwellr> mgmt net means different things to different people
15:33:32 <tidwellr> but it's typically a mgmt net of sorts
15:33:40 <baoli> tidwellr, got it, thanks
15:34:00 <carl_baldwin> #topic IPAM
15:34:08 <carl_baldwin> Pluggable IPAM is all merged!
15:34:22 <johnbelamaric> Yes, indeed. :)
15:34:23 <pavel_bondar> great!
15:34:27 <pavel_bondar> :)
15:34:32 <carl_baldwin> johnbelamaric: pavel_bondar:  Anything for today?
15:34:37 <regXboi> the eagle has landed!!!!
15:34:59 <johnbelamaric> We are working on our driver and have created networking-infoblox repo - but i am having trouble getting attention from anyone in infra to add me to the core/release groups for it
15:35:02 <carl_baldwin> From my perspective, we’ve got the switch over to pluggable by default in Mitaka.  Any other future directions?
15:35:03 <johnbelamaric> any suggestions?
15:35:27 <carl_baldwin> johnbelamaric: Hmmm.  I’m not sure how to get that kind of attention.  mestery ?
15:35:45 <johnbelamaric> No, not for this - i think the migration for Mitaka and then it's a wrap.
15:35:55 <mestery> johnbelamaric: Let me help with this, thanks for tagging me here carl_baldwin
15:36:00 <mestery> johnbelamaric: We'll get it sorted ASAP
15:36:02 <johnbelamaric> mestery: thank you
15:36:09 <mestery> johnbelamaric: Do you happen to have a link to the gerrit group for me?
15:36:13 <mestery> That would be awesome if so :)
15:36:22 <carl_baldwin> On another note, I was just wondering yesterday about DHCP relays.  But, we can talk about that another time.  johnbelamaric:  Any plans around this?
15:36:59 <johnbelamaric> mestery: #link https://review.openstack.org/#/admin/groups/1009,members
15:37:17 <johnbelamaric> mestery: and #link https://review.openstack.org/#/admin/groups/1010,members
15:37:39 <johnbelamaric> carl_baldwin: well, we still have the issue that the ISC patch we need is not upstream yet - i can check into that
15:38:06 <carl_baldwin> johnbelamaric: It is all coming back to me.  I had forgotton about that ISC patch.
15:38:09 <johnbelamaric> Our current implementation is less than ideal, though, since it runs a relay per namespace
15:38:32 <johnbelamaric> ideally we would have a relay that pokes into all the namespaces to avoid so many processes
15:39:14 <carl_baldwin> johnbelamaric: Maybe we could find some time in Tokyo to chat about  it.
15:39:16 <johnbelamaric> we also still run dnsmasq for dns proxy in our case, so it's adding to the load not offloading - though you avoid the disk I/O associated with dnsmasq config files
15:39:50 <johnbelamaric> carl_baldwin: Ok, let's try - i am about 75% certain I will make Tokyo
15:40:21 <carl_baldwin> johnbelamaric: Thanks.  Once again, thanks for your persistence and patience with pluggable IPAM
15:40:27 <carl_baldwin> pavel_bondar too
15:40:45 <carl_baldwin> #topic DNS
15:40:57 <mlavalle> carl_baldwin: after your review to https://review.openstack.org/#/c/200952/
15:40:58 <carl_baldwin> mlavalle: Hi
15:41:14 <mlavalle> I am feeling that it is close to merger
15:41:40 <mlavalle> i will finish ading ut's today and address your concerns
15:41:51 <carl_baldwin> mlavalle: I think so.
15:41:54 <mlavalle> after that i think it is very close
15:42:19 <carl_baldwin> I won’t be able to review again until next week but hopefully other reviewers can add it to their list.
15:42:19 <mlavalle> so i think statrting tomorrow my focus will shift mainly to the external side of this project
15:42:47 <mlavalle> i hope to be able o push the first patchset middle of next week
15:43:03 <mlavalle> so all in all making good progress
15:43:04 <carl_baldwin> mlavalle: That is great.  I plan to run the patch in devstack next week.  This was my first blueprint, I’m excited to see it working.
15:43:28 <mlavalle> thansks to regXboi, who keeps revieweing each one of my revisions :-)
15:43:45 <regXboi> mlavalle: yw
15:43:46 <carl_baldwin> Request reviewers to watch for an update to this patch in the next day.
15:44:04 <mlavalle> yeah, i'll push tonight again
15:44:07 <carl_baldwin> regXboi is all over the place!
15:44:23 <carl_baldwin> mlavalle: Anything else?
15:44:28 <mlavalle> that's all for today
15:44:35 <regXboi> carl_baldwin: I think I've been accused of being an AI because of that
15:45:09 * carl_baldwin will look in to adding captchas to regXboi ’s reviews.
15:45:13 <john-davidge> regXboi: That’s exactly what an AI would say…
15:45:34 <regXboi> rotfl
15:45:43 <regXboi> folks - ask mestery - he know's I'm real
15:45:57 <regXboi> he's watched me throw reading glasses across a room in a fit of pique
15:45:59 <mestery> lol
15:46:04 <mestery> I can confirm regXboi is indeed an actual person
15:46:06 <carl_baldwin> mestery might be an AI too. 
15:46:15 <regXboi> rotflmao now
15:46:21 <mestery> rofl
15:46:27 <mlavalle> i've suspected that of mestery for a long time
15:46:30 <carl_baldwin> #topic Address Scopes
15:46:35 <mestery> rofl
15:46:44 <carl_baldwin> rofl too
15:47:24 * john-davidge suggests adding Turing Test to OpenStack contributors agreement
15:47:25 <carl_baldwin> So, address scopes.  The l3 agent code is mostly finished in a local sandbox.  That was my story three weeks ago, I know.
15:48:06 <carl_baldwin> But, I’m going to wrap this up before moving on to routing networks redesign next week.
15:48:18 <carl_baldwin> vikram_ ’s and numan ’s patches have been making good progress.
15:48:23 <numan> Caleb--, there is one review - to associate subnetpool to addresscope : https://review.openstack.org/#/c/197552/
15:48:27 <numan> carl_baldwin, ^
15:48:28 * regXboi suggests adding Turing Test to OpenStack contributions workflow
15:48:42 <numan> just need another core reviewer to review :)
15:49:15 <carl_baldwin> haleyb: ^
15:49:43 <carl_baldwin> ZZelle is on vacation by the way.  Otherwise, I would’ve welcomed him as our newest core earlier in the meeting.
15:49:56 <haleyb> carl_baldwin: will look
15:50:19 <carl_baldwin> thanks
15:50:29 <carl_baldwin> #topic IPv6
15:50:49 <carl_baldwin> john-davidge: Thank you for your tireless work on prefix delegation.
15:51:06 <carl_baldwin> The first patch merged.
15:51:31 <john-davidge> carl_baldwin: Thank you for your tireless reviewing on prefix delegation
15:51:37 <john-davidge> HenryG: Too
15:52:16 <john-davidge> I actually have a quick fix patch up for an error I didn’t catch until testing the second patch this morning #link https://review.openstack.org/#/c/209978/
15:52:36 <john-davidge> It’s very small, so hopefully it can merge quickly
15:52:54 <carl_baldwin> The next one is a big one but it is mostly new code with what looks like light integration with the existing code.  So, I’m not sure it would make sense to break it up.
15:53:37 <john-davidge> carl_baldwin: Thanks for your review on #link https://review.openstack.org/#/c/185977/ already. I’m working through it now, will possibly have a new patch up sometime tomorrow
15:53:38 * carl_baldwin is a bit embarassed by the that bug.  I thought I checked for that specifically in my last review.
15:54:06 * HenryG is equally embarassed
15:54:11 <carl_baldwin> john-davidge: I won’t be able to review until late next week.  Encourage other reviewers to add it.
15:54:12 <john-davidge> carl_baldwin: Me too! I suspect a bit of “Let’s get this in already” fever caught us there
15:54:25 <john-davidge> carl_baldwin: will do
15:55:10 <john-davidge> carl_baldwin: I agree that it doesn’t make sense to split the second patch. I did try it at one point and it was a lot of overhead
15:55:34 <carl_baldwin> john-davidge: agreed
15:55:45 <HenryG> ack
15:56:09 <carl_baldwin> john-davidge: Do you think a UT could be added for the bug?
15:56:52 <john-davidge> carl_baldwin: Could do. iirc there are no existing tests for update_subnet so I didn’t initally see the value in creating one just for PD
15:57:26 <carl_baldwin> ouch!
15:57:38 <john-davidge> carl_baldwin: In this case it’s a misnamed function call rather than a logical error, so perhaps a UT is a bit much?
15:58:02 <regXboi> um
15:58:05 <carl_baldwin> HenryG: thoughts?
15:58:29 <carl_baldwin> I think it is worth a shot.  UTs in python are useful for catching such things.
15:58:37 * regXboi admits to being somewhat disturbed that there are no tests for update_subnet
15:58:47 <HenryG> It might be worth throwing together a minimal test for update_subnet
15:59:01 <carl_baldwin> 1+1
15:59:03 <carl_baldwin> +1
15:59:10 <john-davidge> carl_baldwin: As in, perhaps a test for that one function is a bit much. But testing of the workflow in #link https://review.openstack.org/#/c/185977/ should absolutely catch it.
15:59:15 <carl_baldwin> == 2 lol
16:00:06 <regXboi> wow, where has the time gone?
16:00:07 <HenryG> Also wondering: do we functional testing for dibbler? I am not sure how to emulate a PD server.
16:00:12 <carl_baldwin> We need to end the meeting now.
16:00:17 <carl_baldwin> We’ll take the discussion to the review.
16:00:24 <carl_baldwin> Thanks, all!  Great work!
16:00:26 <john-davidge> okey dokey
16:00:26 <regXboi> oom
16:01:49 <regXboi> carl_baldwin: #endmeeting?
16:01:52 <numan> bye
16:02:18 <carl_baldwin> #endmeeting