15:00:58 <haleyb> #startmeeting neutron_dvr
15:00:58 <openstack> Meeting started Wed Apr  5 15:00:58 2017 UTC and is due to finish in 60 minutes.  The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:02 <openstack> The meeting name has been set to 'neutron_dvr'
15:01:19 <haleyb> #chair Swami
15:01:20 <openstack> Current chairs: Swami haleyb
15:01:38 <haleyb> #topic Announcements
15:02:06 <Swami> haleyb: sorry I was not able to join last weeks meeting, since I need to take my son to the Dr's office.
15:02:29 <haleyb> Swami: np, i actually cancelled as i had a conflict
15:02:47 <haleyb> hopefully he is feeling better
15:02:49 <Swami> ok,
15:02:58 <Swami> haleyb: yes lot of spring allergies
15:03:29 <haleyb> that's better than the colds going around here :-(
15:03:37 <haleyb> anyways....
15:03:46 <haleyb> Pike-1 is next week, https://releases.openstack.org/pike/schedule.html
15:04:23 <Swami> haleyb: hope you are out of the blizzard from last week.
15:05:15 <haleyb> blizzard was only a few inches here, other places got a lot more, snow is finally starting to melt
15:05:39 <Swami> haleyb: I did see that you have approved the server side patch for the fast path exit. May be both the server side and agent side we can target for Pike milestone1
15:06:22 <haleyb> other announcement is Summit is in one month, so we need to start working on that presentation :)
15:06:51 <haleyb> Swami: yes, i just went through that again this morning, will get to the other one after the meeting
15:07:37 <haleyb> we need to find another reviewer familiar with the plan, i would ping oleg, just don't see him online
15:07:49 <haleyb> s/plan/RFEs
15:07:53 <haleyb> and code
15:08:09 <Swami> haleyb: ok makes sense
15:08:47 <Swami> haleyb: #link https://review.openstack.org/#/c/434863/
15:08:55 <haleyb> of course we should ping mlavalle as well, don't see him either, but we should keep him in-sync regarding work items
15:09:27 <Swami> quick question on this patch, I have not payed attention to this. But it seems that it kind of reverts my original fix where it cleansup the stale namespace when the agent restarts.
15:09:54 <Swami> haleyb: I will take a closer look at it today.
15:10:15 <haleyb> Swami: is there not an issue when HA+DVR?  let me drop the +2 until you look
15:10:33 <Swami> haleyb: no problem.
15:10:59 <haleyb> done
15:11:00 <Swami> haleyb: I just looked, the fix is simple and only addresses the 'ha' + 'dvr' case. So we are fine.
15:11:21 <Swami> haleyb: we both are too quick
15:11:34 <haleyb> Swami: ok, thanks, i was too quick to remove the +2, didn't want someone to see my approval and merge quickly
15:12:05 <haleyb> i'll wait for the recheck anyways, it's been 7 weeks since it was last run
15:12:19 <haleyb> let's move to bugs if there are no other announcements
15:12:24 <haleyb> #topic Bugs
15:12:30 <Swami> haleyb: I will +1 once the jenkins completes.
15:12:57 <Swami> haleyb: I did not see any new bugs this week.
15:13:26 <Swami> haleyb: Still we have one bug that might need a triage.
15:13:36 <haleyb> I didn't see new ones either
15:13:58 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1657981
15:13:58 <openstack> Launchpad bug 1657981 in neutron "FloatingIPs not reachable after restart of compute node (DVR)" [Undecided,New]
15:14:26 <Swami> This one needs triage, this keeps falling off my radar. But I will test it today. I am stacking right now.
15:15:18 <haleyb> ok. can you test that in devstack? looks like it needs a reboot and devstack doesn't like that
15:15:57 <Swami> haleyb: Let me check, I am not sure if devstack will panic when the agent goes down.
15:16:40 <Swami> haleyb: so what is best way to test it without devstack.
15:16:59 <haleyb> HOS :)
15:17:10 <Swami> haleyb: sure,
15:17:50 <Swami> The next one is the lbaas issue
15:17:53 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1672345
15:17:54 <openstack> Launchpad bug 1672345 in neutron "Loadbalancer V2 ports are not serviced by DVR" [Undecided,New]
15:18:20 <Swami> This we have already discussed and it should be related to the allowed_address_pair port.
15:18:48 <Swami> I do have patch for review for the allowed_address_pair ports.
15:19:17 <Swami> #link
15:19:18 <Swami> https://bugs.launchpad.net/neutron/+bug/1644231
15:19:19 <openstack> Launchpad bug 1644231 in neutron "fip router config is not created if the vm ports attached to FIPs have no device_owner" [Low,Incomplete]
15:19:38 <Swami> This bug has a patch and probably needs another +2.
15:19:53 <Swami> #link https://review.openstack.org/#/c/425919/
15:20:40 <haleyb> i had abandoned that change
15:21:18 <Swami> haleyb: ok, is there any other way that we should document it.
15:22:24 <haleyb> Swami: based on armando's last comment i'm not sure
15:23:08 <Swami> haleyb: ok then let us put on hold for now, if it is not a higher priority
15:24:17 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1612804
15:24:17 <openstack> Launchpad bug 1612804 in neutron "test_shelve_instance fails with sshtimeout" [High,Incomplete]
15:24:47 <Swami> This bug has been marked incomplete, didn't we have it in a monitor state for a while.
15:24:52 <Swami> Can we close this bug
15:25:32 <haleyb> Yes, i will close it and move off our list
15:25:44 <Swami> haleyb: thanks
15:27:10 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1632540
15:27:10 <openstack> Launchpad bug 1632540 in neutron "l3-agent print the ERROR log in l3 log file continuously ,finally fill file space,leading to crash the l3-agent service" [Undecided,In progress] - Assigned to Quan Tian (tianquan23)
15:28:07 <haleyb> that was linked to that HA patch from earlier, but that would just be a partial fix
15:28:31 <Swami> I don't see any new patch for this bug. Yes this is linked to the patch that we just discussed. #link https://review.openstack.org/#/c/434863/
15:28:55 <Swami> haleyb: probably it is a partial fix since it says that logs will be reduced.
15:29:21 <haleyb> right. the other changes proposed just removed the error, which isn't right
15:30:05 <Swami> ok.
15:30:16 <Swami> That's all we have for bugs today.
15:30:23 <Swami> The remaining things are the RFE
15:30:44 <Swami> I think we discussed about the 'fast-path-exit' RFE and you have already reviewed it.
15:31:08 <Swami> #link https://review.openstack.org/#/c/283757/
15:31:21 <Swami> #link https://review.openstack.org/#/c/355062/
15:31:36 <haleyb> yes, will get to second one later
15:32:39 <haleyb> #topic Gate
15:32:48 <Swami> haleyb: I got some initial review comments from Oleg on the allowed_address_pair ports
15:32:51 <Swami> #link https://review.openstack.org/#/c/437970/
15:33:22 <Swami> haleyb: may be you can also review it again to see if there is any issues.
15:33:24 <Swami> #link https://review.openstack.org/#/c/437986/
15:33:52 <haleyb> will do
15:33:54 <Swami> haleyb: oleg's patch that he proposed for the agent side was aborted and I also tested it and was not feasible.
15:34:29 <Swami> haleyb: ok thanks.
15:34:38 <Swami> haleyb: back to the gate issues
15:34:45 <haleyb> ok
15:35:18 <haleyb> In yesterday's CI meeting it was noted a new job was added to the neutron job as non-voting
15:35:28 <haleyb> http://grafana.openstack.org/dashboard/db/neutron-failure-rate
15:35:40 <haleyb> in the dvr check queue section
15:35:55 <haleyb> yesterday it was around 35% failure
15:36:00 <haleyb> let me grab the job
15:38:02 <Swami> haleyb: was it a dvr or ha related job
15:38:14 <haleyb> i'm looking through the meeting logs
15:38:59 <haleyb> https://review.openstack.org/#/c/452294/ was the change
15:39:32 <haleyb> ok, it just moved to the grafana panel, think it's always been there
15:39:57 <haleyb> gate-tempest-dsvm-neutron-dvr-multinode-full-ubuntu-xenial-nv
15:40:47 <haleyb> it seems happier today, so could have been something else
15:41:00 <Swami> ok
15:41:24 <Swami> what does that 'nv' stands for.
15:41:37 <haleyb> either way, looking forward, now that the DVR+HA change merged https://review.openstack.org/#/c/383827/
15:41:42 <haleyb> nv == non-voting
15:42:36 <haleyb> we need to create a job for DVR+HA basically, i think we'll cover in L3 meeting tomorrow, but just wanted to raise it here too
15:42:39 <Swami> ok, cool
15:43:13 <haleyb> that way it will be the multinode job, and the dvr-multinode could eventually go away once it's stable
15:43:32 <haleyb> no need to have a 2-node and 3-node job
15:43:37 <Swami> haleyb: makes sense
15:43:58 <Swami> haleyb: We can have one instance of 3 node job that is sufficient
15:44:32 <haleyb> of course it will require a patch :)  yes, will just have the dvr+ha 3-node job
15:45:28 <haleyb> so that work is on miguel and mine's plates
15:46:13 <Swami> haleyb: great
15:47:09 <haleyb> i still would like to know why the current check jobs for dvr-multinode are so "bad", i.e. the scenario job is at ~99%
15:47:43 <Swami> haleyb: Yes we need to find out the root cause.
15:47:47 <haleyb> i know it's non-voting, but if the job is completely broken and needs a tweak we should help it along
15:48:44 <haleyb> i'll see if it's something simple from a recent run, have no idea
15:49:05 <haleyb> but at this point it's just taking time and resources
15:49:35 <Swami> haleyb: ok let me if I can take a look at it, if I have some time.
15:50:03 <haleyb> Swami: thanks, i'd assume any recent patchset would have it failing
15:51:17 <haleyb> #topic Backports
15:52:06 <Swami> I don't have any backports at this time waiting for review.
15:52:16 <haleyb> https://review.openstack.org/#/c/453663/ and https://review.openstack.org/#/c/453664/ were just created for the stale floating IP configured on rfp device, the master change just merged
15:52:50 <haleyb> that had been lurking for a long time
15:53:23 <haleyb> think that's all that is new there
15:53:25 <Swami> Yes, will push a backport patch for this.
15:53:47 <haleyb> i went back to newton, was probably broken in mitaka as well
15:53:48 <Swami> Sorry it is already in stable/newton.
15:53:53 <Swami> will review it.
15:54:33 <haleyb> but mitaka is really critical only, and if you reboot it will fix itself :)
15:54:55 <haleyb> #topic Open Discussion
15:55:07 <haleyb> we have a few minutes left, anything else to discuss?
15:56:05 <Swami> I don't have anything else. Just to let you know, I have not started the slide deck yet for our presentation. Will start some time this week.
15:56:47 <haleyb> Ok, thanks, let me know and we can work on it together
15:57:16 <Swami> Yes, will let you know once I have the skeleton and we can work on
15:58:04 <haleyb> ok, i'll give you two minutes back
15:58:15 <Swami> ok, thanks
15:58:17 <haleyb> #endmeeting