15:00:46 <mlavalle> #startmeeting neutron_l3
15:00:46 <openstack> Meeting started Thu Sep  7 15:00:46 2017 UTC and is due to finish in 60 minutes.  The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:50 <openstack> The meeting name has been set to 'neutron_l3'
15:00:59 <Swami> hi
15:01:06 <davidsha> Hi
15:01:10 <haleyb> hi
15:01:24 <mlavalle> #chair Swami haleyb davidsha
15:01:25 <openstack> Current chairs: Swami davidsha haleyb mlavalle
15:01:40 <mlavalle> Nice to see you team :-)
15:01:50 <mlavalle> #topic Annoucements
15:02:26 <mlavalle> Next week PTG in Denver. Looking forward to see y'all there
15:02:30 <mlavalle> :)
15:03:31 <mlavalle> Also, I wanted to draw the team's attention to the Queens release schedule:
15:03:36 <Swami> mlavalle: do we have an agenda for the PTG neutron team.
15:03:47 <davidsha> Swami: Etherpad was updated
15:03:52 <mlavalle> #link https://releases.openstack.org/queens/schedule.html
15:03:56 <Swami> davidsha: thanks.
15:04:17 <davidsha> Swami: https://etherpad.openstack.org/p/neutron-queens-ptg
15:04:21 <mlavalle> Next milestone will be Queens 1 October 20 - 26
15:04:44 <mlavalle> Swami: Yes, I worked with kevinbenton yesterday afternoong in an agenda
15:04:56 <Swami> mlavalle: good to know
15:05:48 <davidsha> mlavalle: you mean 16th-20th?
15:05:56 <mlavalle> essentially we swept all the topics that folks put in the etherpad and distributed them in sessions across Wednesday through Friday
15:06:15 <mlavalle> Wednesday and Friday are Neutron days
15:06:58 <mlavalle> and Friday will let small interest groups to gather around stadium projects
15:07:07 <mlavalle> makes sense?
15:07:14 <Swami> mlavalle: thanks for the information.
15:07:35 <davidsha> Will there be some way to dial in?
15:07:53 <davidsha> I'd seen it discussed just not sure what the conclusion is.
15:08:05 <mlavalle> I also heard yesterday that the Cinder team is trying to enable a hangout that creates you tube video
15:08:30 <mlavalle> so I'll reach out to them to see if we can do the same thing for Neutron
15:08:43 <davidsha> Thanks!
15:09:08 <mlavalle> that way folks can attend virtualy on their own time
15:10:02 <Swami> mlavalle: yes that would work for people who are not able to attend in person.
15:10:41 <mlavalle> Finally, I wanted to remind folks that the Sydney Summit is fast approaching, November 6 - 8
15:11:02 <mlavalle> You are probably going to need a visa if you plan to attend
15:11:40 <Swami> mlavalle: I got my approval yesterday and will start my visa process.
15:11:47 <mlavalle> For US passport holders (and many other nationalitites) it's just a matter of applying on line and paying about $15 US. Yu get the visa in 5 minutes
15:12:05 <mlavalle> everything on line
15:13:01 <Swami> mlavalle: yeah that should be faster.
15:13:19 <mlavalle> Last two announcements: we will have a Neutron social event in Denver Thursday night. Hopefully all of you will be there
15:13:59 <mlavalle> and the very last one: carl_baldwin said to me a few weeks ago that he wants to join as many as possible of us for dinner on Monday or Tuesday
15:14:10 <mlavalle> I am pinging him today to firm up the day
15:14:26 <mlavalle> I'll keep you posted
15:15:13 <mlavalle> Any other announcements?
15:15:31 <Swami> mlavalle: nothing from me.
15:16:03 <mlavalle> ok moving on
15:16:12 <mlavalle> #topic Bugs
15:16:26 <mlavalle> as usual, Swami plase fire away
15:16:29 <Swami> mlavalle: thanks
15:16:42 <Swami> For DVR we have a critical bug and work is in progress.
15:16:59 <Swami> We have two patches so far pushed in for review for this bug.
15:17:22 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1713927
15:17:23 <openstack> Launchpad bug 1713927 in neutron "gate-grenade-dsvm-neutron-dvr-multinode-ubuntu-xenial fails constantly" [Critical,In progress] - Assigned to Swaminathan Vasudevan (swaminathan-vasudevan)
15:17:42 <Swami> The agent side patch is to prevent agent processing a fip that does not have a host key.
15:18:06 <Swami> #link https://review.openstack.org/#/c/499725/
15:18:36 <Swami> This patch has already been reviewed and had a workflow, but the gate is failing jenkins. I have applied a recheck on it.
15:19:03 <Swami> The server side patch is here #link https://review.openstack.org/#/c/499859/
15:19:36 <haleyb> Swami: i also updated https://review.openstack.org/#/c/500143/ - should work the same as yours, was using it for debugging here
15:19:54 <Swami> I did some testing with this patch, but I found that the multinode grenade job is still failing. I need to look at the logs, to see if this is due to the current change or something else.
15:20:04 <mlavalle> so these last two patchsets do the same thing?
15:20:21 <haleyb> mlavalle: well, they should in theory
15:20:37 <Swami> haleyb: thanks.
15:20:53 <mlavalle> but we are merging only one of them, right?
15:21:08 <haleyb> the goal when the agent and server patch are merged is the multinode job should be passing
15:21:17 <Swami> haleyb: mlavalle: I will take a look at the logs again today and see why it fails in tempest 'Networkscenario' tests.
15:21:37 <haleyb> mlavalle: yes, since it was a hot topic we were both looking at failure cases
15:21:58 <Swami> haleyb: We are not sure there are some other traces also seen in the multinode job.
15:22:47 <haleyb> Swami: we will need to look, there shouldn't be any traces from our changes
15:23:04 <Swami> haleyb: Not from our traces. But it is not related to our change.
15:23:24 <Swami> haleyb: I see an error on 'Segment' and some 'StaleDataError' produced by DB updates.
15:23:28 <haleyb> i was seeing other failures yesterday unrelated to neutron, was hoping a recent devstack-gate fix would have helped
15:23:56 <haleyb> Swami: we should check if there's an existing bug for that
15:24:05 <Swami> haleyb: Sure will do.
15:24:37 <Swami> haleyb: mlavalle: Also a general question, I have issues with running the tempest service with devstack. It fails with 'flake' issue. Have anyone seen this.
15:24:57 <mlavalle> No, I haven't
15:25:09 <haleyb> i haven't seen that either
15:25:25 <Swami> I am having it for a while, ok I will check it out later.
15:25:32 <Swami> Let us move on to the next bugs.
15:25:41 <mlavalle> I have a question
15:26:01 <Swami> mlavalle: yes
15:26:23 <mlavalle> do we need to wait for the approved patchsetto merge before we push either of the other two?
15:26:54 <mlavalle> I just don't want to do something stupid trying to help :-)
15:27:15 <haleyb> mlavalle: i did tell ihar i would be careful :)
15:27:32 <haleyb> kevin pushed the agent patch, which failed the gate anyways
15:27:41 <Swami> mlavalle: what are the approved patches.
15:28:15 <mlavalle> https://review.openstack.org/#/c/499725
15:29:31 <Swami> haleyb: Theoretically this is an agent side patch and can merge without any issues, but as haleyb pointed out it is currently failing merge in gate. Not sure what is going on.
15:30:23 <haleyb> mlavalle: i was going to wait until today to let you look at that, it's a temp fix until we can get the server code working better
15:31:00 <Swami> agreed
15:31:05 <Swami> Let us move on.
15:31:10 <haleyb> Swami: we shouldn't technically need that once we fix the server, but I guess i'd rather have the agent check as well
15:31:11 <Swami> The next bug in the list is
15:31:48 <Swami> haleyb: yes I dont' want to remove it for now. This code path will be hit, only when the server misbehaves. So let us leave it for now.
15:32:03 <haleyb> right
15:32:13 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1714251
15:32:14 <openstack> Launchpad bug 1714251 in neutron "router_centralized_snat not removed when router migrated from DVR to HA" [Undecided,In progress] - Assigned to venkata anil (anil-venkata)
15:32:53 <Swami> This bug is related to the csnat port not getting cleared after router migration. This is also seen tempest testing occassionally.
15:33:04 <Swami> anilvenkata has pushed in a patch for review.
15:33:23 <Swami> #link https://review.openstack.org/#/c/501717/
15:33:30 <Swami> This patch needs review.
15:33:58 <mlavalle> can we review and merge regardless of the other work you and haleyb are doing?
15:34:13 <Swami> Right now jenkins is not happy with this patch.
15:34:17 * mlavalle is suffering a mild case of PTSD
15:34:41 <Swami> mlavalle: Sure if there is any patch on our codepath we will let you know. The work we are doing is on the full sync path.
15:34:51 <mlavalle> ok
15:34:54 <haleyb> mlavalle: yes, there are a few migration fixes that can merge when good, the server one we need to triple-check
15:34:59 <davidsha> jli
15:35:51 <Swami> The next bug is #link https://bugs.launchpad.net/neutron/+bug/1714796
15:35:52 <openstack> Launchpad bug 1714796 in neutron "Router interface not updated during DVR->HA, HA->DVR migratons" [Undecided,In progress] - Assigned to venkata anil (anil-venkata)
15:36:10 <Swami> There is also a patch to address this issue.
15:36:38 <Swami> #link https://review.openstack.org/500379
15:36:55 <Swami> The bug states that the device_owner is not updated during the migration.
15:37:07 <Swami> I saw haleyb had already reviewed this patch.
15:37:20 <Swami> I will include it into my review list as well
15:38:25 <Swami> The next bug is also related to migration with the combination of HA and DVR
15:38:28 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1714802
15:38:29 <openstack> Launchpad bug 1714802 in neutron "DVR and HA migration tests failing intermittently for gate-tempest-dsvm-neutron-dvr-multinode-scenario-ubuntu-xenial-nv job" [Undecided,In progress] - Assigned to Brian Haley (brian-haley)
15:38:47 <Swami> anilvenkata has done some work on triaging all these bugs.
15:38:53 <Swami> He also have a patch for it for review.
15:39:13 <Swami> #link https://review.openstack.org/500384
15:39:25 <Swami> Need reviews on these patches.
15:39:36 <haleyb> the plan is to probably split that one into two reviews from what i remember
15:39:59 <Swami> haleyb: ok
15:40:07 <haleyb> he also noticed the segment issues, let me see if he filed a bug
15:40:07 <Swami> The next on is #link https://bugs.launchpad.net/neutron/+bug/1709774
15:40:09 <openstack> Launchpad bug 1709774 in neutron "Multiple router_centralized_snat interfaces created during Heat deployment" [High,In progress] - Assigned to Swaminathan Vasudevan (swaminathan-vasudevan)
15:40:41 <Swami> haleyb: I did not see one tagged with DVR on the segment issue, but it might be ml2 specific bug.
15:41:16 <Swami> #link https://review.openstack.org/#/c/494376/
15:41:47 <Swami> This patch needs review. mlavalle had already reviewed this patch a couple of times. May be we need to take a look at it again.
15:42:09 <Swami> This patch is can merge on its own and will not cause any trouble to our existing critical issue in the gate.
15:42:41 <Swami> That's all I have for today on DVR bugs.
15:42:46 <Swami> mlavalle: back to you.
15:42:52 <haleyb> there was another related to fwaas
15:43:00 <haleyb> https://bugs.launchpad.net/neutron/+bug/1715395
15:43:01 <openstack> Launchpad bug 1715395 in neutron "FWaaS: Firewall creation fails in case of distributed routers (Pike)" [High,In progress] - Assigned to Reedip (reedip-banerjee)
15:43:03 <mlavalle> Swami: I have a question regarding this last patchset
15:43:45 <haleyb> mlavalle: go ahead
15:44:51 <mlavalle> if the router port fails, we want to delete the snat port, correct?
15:45:38 <mlavalle> nevermind, I think the answer is yes
15:45:47 <mlavalle> I just got confused
15:45:55 <Swami> the snat port should be removed when we remove the router-interface. Or in other words, if something errors out while router port is getting created, then we need to also clean up the csnat ports.
15:46:13 <mlavalle> yeap
15:46:18 <Swami> provided the gateway is set.
15:46:22 <mlavalle> and that is why we need the context
15:46:46 <Swami> haleyb: thanks for the buglink on the fwaas. This was not tagged to 'l3_dvr_backlog'. I just tagged it.
15:46:53 <Swami> I already see a patch in progress.
15:47:15 <Swami> #link https://review.openstack.org/501570
15:47:18 <haleyb> yes, i didn't know they were using router_info data like that
15:48:08 <Swami> I will review it. They are applying the firewall rules on the 'rfp' port, so they need the info on the fipnamespace creation and the link.
15:48:26 <haleyb> let's go to mlavalle since he must have bugs as well, only :12 left
15:48:38 <mlavalle> no, I don't have any
15:48:39 <Swami> mlavalle: back to you.
15:48:56 <mlavalle> do we need to discuss something else?
15:49:20 <davidsha> https://bugs.launchpad.net/neutron/+bug/1705536 was discussed at the drivers meeting last week.
15:49:21 <openstack> Launchpad bug 1705536 in neutron "[RFE] L3-agent agent-mode dvr bridge." [Wishlist,Triaged]
15:49:30 <mlavalle> let me change topics
15:49:35 <davidsha> Sorry!
15:49:55 <mlavalle> #topic L3-agent agent-mode dvr bridge
15:49:58 <mlavalle> go ahead
15:50:27 <mlavalle> it was actually very well received, I might say
15:50:32 <davidsha> The logs: http://eavesdrop.openstack.org/meetings/neutron_drivers/2017/neutron_drivers.2017-08-31-22.03.log.html#l-147
15:52:01 <davidsha> Yes, They were discussing it being an experimental job, it was conditional that it was kept isolated from existing DvR features.
15:52:17 <davidsha> mlavalle: that about right?
15:52:17 <Swami> davidsha: understood
15:52:26 <mlavalle> Yes
15:53:21 <mlavalle> so I think the question / ask to Swami and haleyb would be to take a look and make sure that is indded the case, that the DVR code is isolated from this patchset
15:53:22 <davidsha> They also were wondering how it would work with HA and VRR
15:53:41 <Swami> mlavalle: sure will take a look at it.
15:53:49 <davidsha> Swami: Thank you.
15:53:51 <Swami> davidsha: Do we have a spec for it.
15:54:14 <davidsha> Swami: No, but I can draft one if necessary.
15:54:53 <Swami> davidsha: if you can draft one that would be great and I can take a look at it during the PTG for review.
15:55:23 <davidsha> Swami: Ack, I'll work on that asap.
15:56:37 <mlavalle> davidsha: anything else?
15:57:28 <davidsha> mlavalle: nothing off the top of my head, if you have any questions for me I'm on IRC most of the time
15:57:37 <mlavalle> ok
15:58:07 <haleyb> Swami: was this the segment trace you saw?  https://bugs.launchpad.net/bugs/1714068 - if so guess i need to get to fixing it :)
15:58:09 <openstack> Launchpad bug 1714068 in neutron "AttributeError in get_device_details when segment=None" [Medium,New]
15:58:33 <mlavalle> Swami, haleyb there is a bug that I would like to discuss with you. But since we are running out of time, let's do this in the Neutron channel
15:58:44 <Swami> mlavalle: ok will do
15:59:03 <Swami> haleyb: I did not remember seeing the Nonetype: attribute error. But let me recheck
15:59:08 <haleyb> mlavalle: sure, i'll be in a meeting but will try
15:59:29 <mlavalle> ok
15:59:42 <mlavalle> #endmeeting