15:00:45 <mlavalle> #startmeeting neutron_l3
15:00:46 <openstack> Meeting started Thu Sep  6 15:00:45 2018 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'
15:01:05 <Swami> hi
15:01:10 <mlavalle> hi Swami
15:01:15 <tidwellr> hi
15:01:27 <mlavalle> hey tidwellr , nice to see you!
15:01:51 <tidwellr> mlavalle: I hope to be back more often going forward :)
15:02:07 <Swami> mlavalle: tidwellr is back
15:02:14 <mlavalle> tidwellr: that's great! do you have a topic to propose today?
15:02:34 <mlavalle> for discussion today
15:02:36 <mlavalle> ?
15:03:08 <tidwellr> mostly eavesdropping today
15:03:23 <mlavalle> ok, later I will ask you a question
15:03:26 <tidwellr> I do have a brief update on "fast-entrance", but nothing else
15:03:27 <davidsha> Hi
15:03:31 <mlavalle> welcome anyways!!!!
15:03:36 <mlavalle> hi davidsha
15:03:48 <mlavalle> ok let's get going
15:03:54 <mlavalle> #topic Announcements
15:04:31 <mlavalle> Firt, as you all know, the PTG is next week in Denver
15:04:46 <mlavalle> It is my understanding this will be the last PTG
15:04:58 <mlavalle> that is why the Stein cycle is a little longer than normal
15:05:15 <mlavalle> because we are moving everything back to the Summit again
15:05:16 <Swami> mlavalle: Is that good or bad
15:05:39 <mlavalle> I think it is good in that it makes the meetings available to everybody
15:05:41 <Swami> mlavalle: Yes summit is loosing momentum on the openstack nowadays.
15:05:59 <mlavalle> with the current scheme people have to choose
15:06:12 <mlavalle> I now it will be clear that you have two trips a year
15:06:24 <Swami> mlavalle: agreed
15:07:05 <mlavalle> In this etherpad I have schdule all the proposed topics for discussion:  https://etherpad.openstack.org/p/neutron-stein-ptg
15:07:22 <mlavalle> Startring at line 132
15:07:32 <manjeets> so berlin summit still will not have any design sessions ?
15:07:59 <mlavalle> manjeets: I expect it to have some forum session, like the past 2 summits
15:08:10 <mlavalle> but that's really up to us
15:08:20 <mlavalle> if we propose sessions for the forum
15:08:25 <manjeets> ah ok !
15:09:36 <mlavalle> Our next milestone is Stein-1, Oct 22 - 26
15:09:56 <mlavalle> any other announcements?
15:10:21 <mlavalle> ahhh, I have another one....
15:10:33 <mlavalle> Due to the PTG, next week's meeting will be cancelled
15:10:46 <mlavalle> ok, let's move on....
15:10:48 <Swami> mlavalle: Do you have online sessions for PTG
15:11:23 <mlavalle> Swami: I have planned specifically for that..... I can attemot what I did last yeat with the camera, if there is interest
15:11:30 <mlavalle> haven't^^^^
15:12:03 <davidsha> <- interested in a live stream :)
15:12:05 <Swami> mlavalle: ok sure. I will be intereseted in attending remote
15:12:12 * tidwellr is interested as well
15:12:58 <mlavalle> ok, I'll bring the camera with me and see what I can do.... since you are all ganging up on me.... LOL
15:13:05 <davidsha> :D
15:13:21 <mlavalle> let's move on
15:13:27 <mlavalle> #topic Bugs
15:13:42 <mlavalle> Swami: please, fire away
15:13:47 <Swami> Sure
15:13:51 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1790084
15:13:51 <openstack> Launchpad bug 1790084 in neutron "ERROR pyroute2.netns.nslink" [Undecided,New]
15:14:54 <Swami> This is a new bug that is being filed against the 'DVR_NO_EXTERNAL' agent mode. It seems that there is an error log in the logs seen when it tries to remove the namespace. It seems that the namespace delete is going to a wrong node. ( compute node in this case ). I will triage it further and update the bug.
15:15:24 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1786272
15:15:24 <openstack> Launchpad bug 1786272 in neutron "Connection between two virtual routers does not work with DVR" [Medium,In progress] - Assigned to Slawek Kaplonski (slaweq)
15:15:59 <Swami> There is a wip patch. #link https://review.openstack.org/#/c/597567/
15:16:06 * haleyb will lurk, was in downstream meeting
15:16:26 <Swami> i have been testing this patch actively, but still there is some issue with this patch and I am updating my review comments on this patch.
15:16:34 <slaweq> hi
15:16:44 <njohnston> hi
15:16:51 <slaweq> Swami: I will push new PS for it in few minutes
15:16:55 <slaweq> can You test this one?
15:17:10 <Swami> slaweq: sure, I tested the one till yesterday
15:17:29 <Swami> I did see that the routers that are not supposed to be in the compute node was also started, so I was debugging that.
15:17:39 <Swami> slaweq: But I will test your update today as well.
15:17:49 <slaweq> ahh, that I didn't test yet
15:18:09 <slaweq> I was currently focused on testing if case described in bug works with it
15:18:20 <slaweq> and at least this looks that is working for me fine
15:18:39 <Swami> ok no problem, my test case was there was an additional router and I did see that router namespace also poped up on all compute nodes.
15:18:50 <slaweq> but there can be other "side effects" which I will have to fix still
15:19:05 <slaweq> thx for info, I will check that also
15:19:20 <Swami> slaweq: Good work, so far. Yes it solves the bug description. But it tries to create all router namespace now.
15:19:34 <slaweq> also, I pushed patch with scenario test to cover this: https://review.openstack.org/#/c/598676/
15:19:40 <Swami> slaweq: we should probably filter only based on the connected routers and not all routers.
15:19:44 <slaweq> please take a look if You will have time
15:19:57 <Swami> slaweq: Yes I did see that test patch passed
15:20:13 <slaweq> Swami: ok, thx for feedback and testing it
15:20:24 <Swami> slaweq: I will test your current patch and will post my reviews.
15:20:32 <slaweq> I will continue work on it and try to fix this issue with additional routers
15:20:38 <slaweq> thx Swami :)
15:20:43 <Swami> slaweq: thanks for your help.
15:20:50 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1774459
15:20:50 <openstack> Launchpad bug 1774459 in neutron "Update permanent ARP entries for allowed_address_pair IPs in DVR Routers" [High,Confirmed]
15:21:19 <Swami> mlavalle: slaweq: I need some l2 experts to take a look at the proposal. I will try to post a WIP patch for this.
15:21:38 <Swami> Right now I am having some trouble in forwarding the packet to the 'tap' interface.
15:22:19 <mlavalle> haleyb proposed this topic for discussion in the PTG. There is a session, L3 topics, on Thursday, weher this is scheduled to be discussed
15:22:30 <Swami> I will post a wip patch and then we can discuss about the forward motion on this. In the mean time if there is any other efficient way of tapping the GARP packets, that would help me speed up the process.
15:22:42 <Swami> mlavalle: Sure then it would be great.
15:23:07 <mlavalle> so if you leave as much feedback in the bug as possible, that will enrich the conversation
15:23:10 <mlavalle> hang on....
15:23:18 <Swami> mlavalle: Sure, I will do it.
15:24:30 <Swami> That's all I had for today with respect to the DVR bugs.
15:24:36 <mlavalle> Swami: that session is split in two, before and after lunch.... The first half is from 11:45 to 12:15, which will be 11:45 to 11:15 your time. Is that a good time for you?
15:24:58 <mlavalle> since we are going to try to do remote session, you might as well participate
15:25:24 <Swami> mlavalle: That should be good. I can join.
15:25:36 <mlavalle> I maenat 10:45 to 11:15 your time
15:25:39 <Swami> mlavalle: yes if the remote session is on I will join.
15:25:52 <Swami> mlavalle: no problem.
15:26:09 <mlavalle> I still advise to leave as much feedback in the bug itself, in case we have technical difficulties
15:26:30 <Swami> mlavalle: sure I will do as much as possible.
15:26:34 <mlavalle> you know, stupid computers not always do what you want ;-)
15:27:20 <Swami> mlavalle: understood.
15:27:30 <Swami> mlavalle: that's all I have.
15:27:42 <mlavalle> Thanks Swami
15:28:12 <mlavalle> I have the following bugs....
15:28:20 <mlavalle> First https://bugs.launchpad.net/neutron/+bug/1788006
15:28:20 <openstack> Launchpad bug 1788006 in neutron "Tests fail with "Server [UUID] failed to reach ACTIVE status and task state "None" within the required time ([INTEGER] s). Current status: BUILD. Current task state: spawning."" [Critical,Confirmed] - Assigned to Miguel Lavalle (minsel)
15:28:38 <mlavalle> for this one we have concluded that we need help from the Nova team
15:28:58 <mlavalle> slaweq and I talked to Matt a couple days ago and he said he will help out
15:29:13 <mlavalle> so on hold for the time being on this one
15:29:31 <mlavalle> Next one is https://bugs.launchpad.net/neutron/+bug/1789434
15:29:31 <openstack> Launchpad bug 1789434 in neutron "neutron_tempest_plugin.scenario.test_migration.NetworkMigrationFromHA failing 100% times" [High,Confirmed] - Assigned to Miguel Lavalle (minsel)
15:30:32 <mlavalle> the problem here is that we have a scenario test to migrate from HA to {legacy | DVR | HA-DVR}
15:31:18 <mlavalle> the test assumes that after settin the router admin_state_up to False, prior to the migration, all the router ports go down
15:31:45 <mlavalle> haleyb did some testing a couple of days ago, that conflicts with my result of yesterday
15:31:50 <mlavalle> let me explain....
15:32:11 <mlavalle> I start with a HA router, excatly like the scenario test
15:32:28 <mlavalle> when I set admin-state_up to False, all the ports state active
15:32:56 <mlavalle> haleyb: in you case, I think I understoof you started with a DVR router.... Is my understanding correct?
15:33:16 <haleyb> mlavalle: yes, it was dvr originally
15:33:41 <mlavalle> haleyb: ok, I will test that case and see if I get the same result as you
15:33:48 <haleyb> brought down, changed ha, now it's in weird state regarding ports always active
15:34:32 <mlavalle> haleyb: ahhh, maybe, once you made it HA, you are in effect testing the same thing as I did
15:35:49 <haleyb> maybe, yes
15:36:12 <mlavalle> I'll play with these different paths
15:37:28 <mlavalle> now, what should it be haleyb and Swami? when I set a router (legacy, DVR, HA, HA-DVR) admin_state_down to False, all its ports should become inactive?
15:38:19 <haleyb> i think so
15:39:24 <mlavalle> yes, that is my thinking as well, just making sure we are all on the same page
15:39:50 <Swami> mlavalle: I agree
15:39:56 <mlavalle> cool
15:40:02 <slaweq> it was like that when I was writing this check some time ago - when router was set to admin_state_up=False, all it's ports were set to be DOWN
15:40:37 <mlavalle> The last one I have is https://bugs.launchpad.net/neutron/+bug/1787919
15:40:37 <openstack> Launchpad bug 1787919 in neutron "Upgrade router to L3 HA broke IPv6" [High,Confirmed] - Assigned to Miguel Lavalle (minsel)
15:41:48 <mlavalle> The submitter in note #10 proposes what the solution should be in his opinion.... before going down that way, I would like Swami and haleyb to chime in
15:41:52 <mlavalle> would you?
15:42:14 <mlavalle> just leave a comment there when you have some time
15:42:19 <haleyb> ack
15:42:19 <Swami> mlavalle: I will take a look at it.
15:43:54 <mlavalle> ok, that's all I have for today in terms of bugs
15:44:07 <mlavalle> any other bugs we should discuss today?
15:45:16 <mlavalle> ok, moving on
15:45:25 <mlavalle> #topic Openflow DVR
15:45:44 <mlavalle> davidsha: since you are here, let me ask a couple of questions....
15:45:55 <davidsha> Sure, go ahead
15:46:53 <mlavalle> First: as discussed a few weeks ago I created a blueprint, https://blueprints.launchpad.net/neutron/+spec/openflow-based-dvr, that I added to our Stein dashboard in https://launchpad.net/neutron/+milestone/stein-1
15:47:07 <davidsha> For a quick update We're working on rebasing the openflow_dvr POC ontop of the refactor
15:47:31 <mlavalle> haleyb accepted to be the approver. Who should be the assignee?
15:47:32 <davidsha> Thank you!
15:47:44 <davidsha> Xubo should be the assignee
15:48:19 <davidsha> I'm limited in how I can contribute but I'll help as much as I can
15:49:12 <mlavalle> so if I type xubo in Launchpad, I get several options. It is diffcult to know which one
15:49:39 <mlavalle> ahhh hang on////
15:49:48 <mlavalle> would it be bo.b.xu@intel.com?
15:50:07 <davidsha> xubo.zhang@intel.com is his email I believe, give me 1 sec to double check
15:50:52 <davidsha> ya it's xubo.zhang@intel.com
15:51:04 <mlavalle> ok, I'll try to find that
15:51:44 <mlavalle> and the second question you answered. should we wait for that rebasing? or it has been accomplished?
15:52:16 <davidsha> It's in progress, the rough rebase is done it just needs to be tested to make sure we didn't miss something
15:52:32 <mlavalle> ok, so we will keep an eye for it
15:52:34 <mlavalle> Thanks!
15:52:43 <davidsha> thank you!
15:53:39 <mlavalle> #topic On demand agenda
15:54:18 <mlavalle> tidwellr: this is where I wanted to ask you.... could you take a look at this RFE and help us triage it: https://bugs.launchpad.net/neutron/+bug/1784590?
15:54:19 <openstack> Launchpad bug 1784590 in neutron "neutron-dynamic-routing bgp agent should have options for MP-BGP" [Wishlist,New]
15:54:32 <tidwellr> sure
15:54:41 <mlavalle> cool, thanks!
15:54:50 <mlavalle> any other topics we should discuss today?
15:55:04 <tidwellr> https://review.openstack.org/#/c/581098/
15:55:34 <tidwellr> I have working code for DVR-aware BGP announcements, this WIP though
15:55:41 * cdent waves at mlavalle
15:55:57 <tidwellr> feel free to take a look and offer feedback
15:55:57 * mlavalle waves back ;-)
15:56:13 <mlavalle> thanks tidwellr. really cool
15:56:31 <njohnston> very cool
15:57:07 <mlavalle> and now looking at the patch, I know what Swami meant by "tidwellr is back". LOL
15:57:13 <Swami> tidwellr: So this patch would completely open up the fast path entry.
15:57:15 <mlavalle> congrats
15:57:18 <mlavalle> !
15:57:30 <Swami> mlavalle: yes I can share my workload with tidwellr now
15:57:39 * tidwellr ducks
15:58:14 <mlavalle> ok guys. Thanks for attending. See you in two weeks....
15:58:23 <Swami> bye.
15:58:25 <davidsha> see you!
15:58:27 <tidwellr> bye
15:58:27 <haleyb> bye
15:58:31 <mlavalle> For those going to Denver, safe travels!
15:58:39 <mlavalle> #endmeeting