15:01:01 <haleyb> #startmeeting neutron_dvr
15:01:02 <openstack> Meeting started Wed Jan 18 15:01:01 2017 UTC and is due to finish in 60 minutes.  The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:06 <openstack> The meeting name has been set to 'neutron_dvr'
15:01:10 <haleyb> #chair Swami
15:01:13 <openstack> Current chairs: Swami haleyb
15:02:25 <jschwarz> hello
15:02:29 <Swami> jschwarz: hi
15:02:36 <haleyb> #topic Announcements
15:02:56 * haleyb had a weird cut/paste issue there for a second
15:03:48 <haleyb> Ocata-3 is iminent
15:04:17 <haleyb> we need to make sure bugs are listed on https://launchpad.net/neutron/+milestone/ocata-3 that we want to land fixes for
15:05:30 <Swami> haleyb: we may have to add couple of DVR bugs to the list I don't see them, like the floatingip static routes, metering bug and also the RFE.
15:05:35 <haleyb> Also, add items to the PTG list at https://etherpad.openstack.org/p/neutron-ptg-pike
15:06:08 <haleyb> Swami: i can add them after the meeting
15:06:20 <Swami> haleyb: thanks
15:07:12 <haleyb> Any other announcements from the team?
15:07:52 <haleyb> #topic Bugs
15:08:19 <Swami> haleyb: thanks
15:08:40 <Swami> haleyb: I did not see any new bugs for this week so far. I am not sure if I missed anything.
15:09:31 <haleyb> Swami: i didn't see any new ones either, unless they were opened this morning
15:09:44 <Swami> jschwarz: do you have any priority bugs that we need to discuss.
15:10:09 <jschwarz> Swami, not a bug but a patch: https://review.openstack.org/#/c/376550/
15:10:34 <jschwarz> haleyb, last week we discussed this and we said that we wanted to do complete tempest coverage for all possible migrations
15:10:45 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1654991
15:10:45 <openstack> Launchpad bug 1654991 in neutron "Update of DVR router to be HA fails with 'HAmodeUpdateOfDvrNotSupported'" [High,In progress] - Assigned to John Schwarz (jschwarz)
15:10:59 <jschwarz> haleyb, however, this seems a bit unlikely for Ocata as it's right upon us, and there are some requirements to be met first
15:11:10 <Swami> jschwarz: yes I remember and you also had a test patch.
15:11:10 <haleyb> jschwarz: yes, since we were removing the check
15:11:10 <jschwarz> specifically, there's no HA gate yet...
15:11:28 <Swami> #link https://review.openstack.org/#/c/418276/
15:11:30 <jschwarz> is it possible to merge the patch and add new tests over time?
15:12:46 <jschwarz> Swami, alas, the test patch is not mine ;-)
15:12:46 <haleyb> jschwarz: do those code paths get tested in a single-node HA setup?
15:12:49 <Swami> jschwarz: You mean new tests in tempest?
15:12:56 <jschwarz> Swami, yes
15:13:22 <jschwarz> haleyb, we don't test them as part of upstream migration process
15:14:02 <jschwarz> haleyb, but we have the DVR gates that check legacy + dvr and once we'll have a HA gate we'll be able to test HA and DVR+HA
15:14:39 <jschwarz> haleyb, in other words, we check connectivity as a whole for legacy and DVR (and in the future, HA)
15:14:49 <jschwarz> just not in a migration
15:19:35 <Swami> jschwarz: we don't support migration of DVR routers to legacy routers I thought.
15:19:35 <jschwarz> quite the net split..
15:19:50 <haleyb> jschwarz: that would be my preference without testing, but that devstack-gate and infra patch could merge soon, right?
15:19:51 <jschwarz> Swami, we aren't - that's what my patch intends to enable, though
15:20:04 <jschwarz> haleyb, could
15:20:16 <haleyb> Swami: what jschwarz said, i think the only remaining issue was vpnaas
15:20:22 <jschwarz> right, so I'll see about getting that infra patch merged
15:20:38 <jschwarz> anilvenkata, ^
15:20:40 <Swami> jschwarz: ok that would be right direction to go forward.
15:21:22 <anilvenkata> haleyb, Swami jschwarz we need Armando's review for that patch
15:22:22 <Swami> anilvenkata: follow up with him or add it to the agenda for the PTG, just to make sure we have something in place to test for ocata.
15:22:34 <anilvenkata> Swami, sure
15:22:43 <anilvenkata> thanks Swami
15:23:47 <Swami> ok, can we move on to the next one
15:23:57 <Swami> jschwarz: do you have anything else.
15:24:09 <jschwarz> Swami, nope :)
15:24:18 <Swami> I did see there is another bug and patch on your name.
15:24:22 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1647432
15:24:44 <jschwarz> Swami, https://review.openstack.org/#/c/407099/
15:24:58 <jschwarz> I need to answer some reviews - hopefully a new revision will be up tomorrow
15:24:59 <Swami> jschwarz: thanks for the patch lin.
15:25:07 <Swami> jschwarz: thanks
15:25:09 <openstack> Launchpad bug 1647432 in neutron "Multiple SIGHUPs to keepalived might trigger re-election" [High,In progress] - Assigned to John Schwarz (jschwarz)
15:25:28 <Swami> The next in the list is
15:25:31 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1644231
15:25:31 <openstack> Launchpad bug 1644231 in neutron "fip router config is not created if the vm ports attached to FIPs have no device_owner" [Low,Triaged]
15:25:44 <jschwarz> is that ajo's?
15:26:01 <Swami> haleyb: I had a question on this. We discussed about having a doc to provide information about the port-binding.
15:26:46 <haleyb> Swami: yes, i think i said i'd take that, and update the docs (whereever they may be)
15:27:11 <Swami> haleyb: I will find a right topic to add this to the documentation. I had some challenge in finding out the right section, since the DVR is only discussed in different implementations and not about the port-binding.
15:27:48 <Swami> Ok we can move on
15:27:49 <haleyb> Swami: it maybe can be in the port create doc, even if it is specific to DVR
15:28:18 <Swami> haleyb: ok
15:28:46 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1644415
15:28:46 <openstack> Launchpad bug 1644415 in neutron "dvr_edge_ha_router disassociates floatingip incompletely" [High,In progress] - Assigned to Zhixin Li (lizhixin)
15:29:00 <Swami> #link https://review.openstack.org/#/c/404571/ - patch link.
15:29:22 <Swami> The owner of the patch have not added any tests. I am working on adding a functional test to this patch. I will post it today.
15:29:28 <Swami> Then we can review it again.
15:29:41 <Swami> Otherwise there is nothing else to discuss in this bug.
15:30:26 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1606741
15:30:26 <openstack> Launchpad bug 1606741 in neutron "Metadata service for instances is unavailable when the l3-agent on the compute host is dvr_snat mode" [Medium,In progress]
15:30:42 <Swami> #link https://review.openstack.org/352686
15:31:45 <haleyb> No udpated in 2 months
15:31:54 <Swami> haleyb: last week I had a action item to fix the merge conflicts and rebase the patch, but then I noticed the review comments on this patch and based on the review comments it seems that we need to come up with a different approach since both carl and assaf had requested to go with different approach.
15:33:05 <Swami> haleyb: need more work on it.
15:33:23 <haleyb> Swami: yes, at this point it's probably not Ocata material
15:33:32 <Swami> haleyb: yes
15:33:47 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1506567
15:33:47 <openstack> Launchpad bug 1506567 in neutron "No information from Neutron Metering agent" [High,In progress] - Assigned to Brian Haley (brian-haley)
15:34:03 <Swami> #link https://review.openstack.org/377108 - patch link
15:34:15 <Swami> Needs review and can be targeted for Ocata-3
15:34:17 <haleyb> I need to review that today
15:35:00 <haleyb> just targeted to o-3
15:35:05 <Swami> haleyb: thanks
15:35:08 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1571676
15:35:08 <openstack> Launchpad bug 1571676 in neutron "After binding a floating IP to VM, the static route can't work in DVR." [Undecided,In progress] - Assigned to Swaminathan Vasudevan (swaminathan-vasudevan)
15:35:22 <Swami> #link https://review.openstack.org/#/c/308068/ - patch link
15:35:33 <Swami> Needs review and also should be targeted for o-3
15:36:16 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1577488
15:36:16 <openstack> Launchpad bug 1577488 in neutron "[RFE]"Fast exit" for compute node egress flows when using DVR" [Wishlist,In progress] - Assigned to Swaminathan Vasudevan (swaminathan-vasudevan)
15:36:30 <Swami> haleyb: this can also be targeted for o-3
15:36:58 <Swami> I did see that there is a test failure in the patch that you uploaded yesterday.
15:37:02 <haleyb> Swami: ok, i'll add those two, don't know if i broke that second one with my update
15:37:09 <Swami> #link https://review.openstack.org/#/c/283757/
15:37:19 <haleyb> it passed locally, haven't looked at logs yet
15:37:27 <Swami> Ok, let me take a look at it.
15:37:49 <Swami> That's all I had for bugs today.
15:38:22 <haleyb> ok, thanks Swami
15:38:32 <haleyb> jschwarz: any HA bugs we missed?
15:38:42 <jschwarz> haleyb, nope :)
15:39:06 <haleyb> #topic Gate
15:39:17 <haleyb> no DVR failures in the gate that i now of :)
15:39:34 <Swami> haleyb: good for the day
15:39:47 <haleyb> still need to update https://review.openstack.org/410973 to make multinode job voting
15:39:55 * haleyb knocks on wood as well
15:40:25 <haleyb> #topic Stable backports
15:40:55 <haleyb> ihar ran a script to find backport candidates, some already proposed, i'll look through remaining bugs
15:41:22 <Swami> I working on couple of backports. I pushed in two patches today. One more to come and I will also see is there any other backports.
15:41:33 <Swami> #link https://review.openstack.org/#/c/421983/
15:41:47 <Swami> #link https://review.openstack.org/#/c/421989/
15:41:53 <haleyb> Swami: yes, feel free to add me, i'll usually show up once jenkins is happy on those
15:42:03 <Swami> haleyb: thanks, will do
15:42:37 <Swami> jschwarz: you mentioned you had a bunch of backports for HA last week, is it done.
15:42:44 <haleyb> jschwarz: you had a series of HA backports i added to the meeting page, don't know if they all mreged, some have
15:43:05 <jschwarz> haleyb, Swami, I'll have to look at that list and report back next week
15:43:16 <Swami> jschwarz: ok no problem.
15:43:20 <jschwarz> I haven't had too much free time this week outside working on a few of my patches
15:43:25 <haleyb> jschwarz: i'll remove or mark the ones that merge since i'm editing the pag already
15:43:39 <jschwarz> haleyb, thanks :)
15:44:24 <haleyb> #topic Open discussion
15:45:01 <jschwarz> nothing from me this week ^_^
15:45:19 <Swami> haleyb: i have a question on IPv6
15:45:44 <haleyb> jschwarz: they all merged, so i will clean up the wiki
15:45:49 <haleyb> Swami: yup
15:45:53 <jschwarz> haleyb, ^_^ ^_^
15:45:55 <jschwarz> awesome!
15:47:07 <haleyb> Swami: still here?
15:47:23 <Swami_> sorry got disconnection
15:47:27 <Swami_> s/disconnection/disconnected
15:47:33 <Swami_> haleyb: I am back
15:48:02 <Swami_> my question is related to fast-path exit if IPv6 address exist.
15:48:47 <Swami_> Today we do have the internal rfp and fpr ports that connect the router namespace to fip namespace and both those ports have a link local address for ipv6.
15:49:14 <haleyb> Swami_: i don't think we can do that yet, since if we did PD the downstream router for the subnet is the centralized one
15:49:50 <Swami_> haleyb: but you mentioned that we can run a partial PD or something from the fip namespace, if I remember right.
15:50:21 <haleyb> if we use subnetpools, we can hack things to make the Fip namespace to proxy ND
15:50:28 <haleyb> s/to/do
15:51:12 <Swami_> haleyb: yes the context is with the subnetpools for the fast-path exit if the subnetpools scope match for ipv6.
15:51:28 <haleyb> Swami_: what we really need is a way to do route injection, that way the fip namespace can be the downstream for a single IPv6 address
15:52:41 <haleyb> Swami_: with a subnetpool we'd just need a little work, there was a patch for the proxy ND work
15:52:59 <Swami_> haleyb: ok will investigate it further. In that case still the rfp and fpr ports can still just have a link-local-address and we should be able to forward the IPv6 traffic to the fipnamespace and get it back, right.
15:54:08 <haleyb> Swami_: i think so, yes.  the review was https://review.openstack.org/#/c/143917/ and it's rather large, but the proxy ND piece i believe is there
15:54:23 <Swami_> haleyb: ok will take a look at it.
15:55:05 <haleyb> i think i've run those 'ip neigh' commands by hand to make it work at least
15:55:13 <Swami_> haleyb: Also for the PTG, one discussion item would be the DVR floatingip with unbound ports. This is pending for a long time and have not seen the tunnel. We need to get a consensus on our approach.
15:55:45 <Swami_> s/seen the tunnel/seen the light of the tunnel
15:55:45 <haleyb> Yes, that would be good to add
15:55:53 <haleyb> i knew what you meant :)
15:56:02 <Swami_> haleyb: thanks
15:56:05 <haleyb> Swami_: are you going tot he PTG?  is anyone else here?
15:56:23 <Swami_> haleyb: I am not sure given the current situation.
15:56:35 <Swami_> haleyb: You should be going.
15:56:53 <Swami_> haleyb: i have not even asked the mgmt yet.
15:56:54 <haleyb> Swami_: i am on vacation that week, Mexico is warmer than Atlanta :)
15:57:19 <Swami_> haleyb: enjoy your vacation then, there is always going to be a PTG report.
15:57:25 <haleyb> jschwarz: are you going?  i'm assuming ihar is
15:57:33 <Swami_> haleyb: hopefully you are going to be at the beach with your beer.
15:57:36 <jschwarz> haleyb, I'm not going
15:57:46 <jschwarz> haleyb, Ihar is indeed going
15:58:15 <haleyb> Swami_: s/beer/beers, and frozen drinks :)
15:58:39 <Swami_> haleyb: :)
15:59:16 <Swami_> haleyb: which place in Mexico?
15:59:20 <haleyb> ok, <1 minute left
15:59:51 <Swami_> haleyb: I don't have anything else.
16:00:11 <haleyb> thanks for attending, let's get those patches merged :)
16:00:15 <haleyb> #endmeeting