15:02:21 <Swami> #startmeeting distributed_virtual_router
15:02:22 <openstack> Meeting started Wed Sep  3 15:02:21 2014 UTC and is due to finish in 60 minutes.  The chair is Swami. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:02:25 <openstack> The meeting name has been set to 'distributed_virtual_router'
15:02:50 <Swami> #info Today will be the Juno 3 cut off date
15:02:59 <Swami> mrsmith: hi
15:03:40 <Swami> So today will be the last day for all Juno features to get in.
15:03:53 <mrsmith> Swami: hiya
15:03:54 <Swami> Rajeev: hi
15:04:02 <Rajeev> Swami: Hi
15:04:10 <Swami> #topic Agenda
15:04:18 <Swami> #topic DVR Update
15:04:59 <Swami> All our DVR patches have landed upstream.
15:05:22 <Swami> We are still working on bugs.
15:06:39 <Swami> #link https://bugs.launchpad.net/neutron/+bugs?field.tag=l3-dvr-backlog
15:06:59 <Swami> Link provides a list of all bugs.
15:08:20 <Swami> Most of the High bugs have assignee's and are worked out right now.
15:08:53 <Swami> There were couple of new bugs that was filed yesterday by Viveknarashimhan
15:09:34 <Rajeev> Swami: one thing to keep in mind, Jenkin queues are backed up so give sufficient time for runs
15:09:39 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1364215
15:09:40 <uvirtbot> Launchpad bug 1364215 in neutron "L2 Agent switch to non-dvr mode on first RPC failure" [Undecided,New]
15:10:07 <viveknarasimhan> i raised that to handle
15:10:11 <Swami> #link https://bugs.launchpad.net/neutron/+bug/1364839
15:10:12 <uvirtbot> Launchpad bug 1364839 in neutron "DVR namespaces not deleted on LBaaS VIP Port removal" [Undecided,In progress]
15:10:23 <viveknarasimhan> the situation discussed in yesterdady meeting
15:10:43 <viveknarasimhan> 1364839 is to address removal LBaaS and DHCP Ports on dvr subnets from service nodes
15:11:44 <Swami> viveknarasimhan: thanks for the update
15:12:34 <Swami> viveknarasimhan: the check_vm_exists_on_host does it not take care of the dvr related ports right now.
15:12:44 <viveknarasimhan> that takes care
15:13:07 <Swami> so what was missing, in there
15:13:10 <viveknarasimhan> but delete_port logic in ml2_plugin does not invoke deletion of namespaces for service pors. it was handling only
15:13:12 <viveknarasimhan> for vm ports
15:13:16 <viveknarasimhan> i posted review here:
15:13:29 <Swami> viveknarasimhan: ok
15:13:59 <Swami> for the 1364215 do we have fix
15:14:09 <Swami> How severe is this bug
15:14:15 <viveknarasimhan> that i am not fixing now
15:14:33 <viveknarasimhan> will fix if we have a consensue that it needs to be fixed
15:14:49 <viveknarasimhan> that was raised as a placeholder for issue seen in TriploO deployment of DVR
15:14:53 <Swami> Ok let us wait till armax triages it
15:15:08 <viveknarasimhan> review link for 1364839 here: https://review.openstack.org/118580
15:15:17 <viveknarasimhan> ok
15:15:34 <Swami> Is this issue seen in devstack or can be reproduced in devstack
15:16:16 <viveknarasimhan> reproducible in devstack also
15:16:20 <viveknarasimhan> if nodes start before controller
15:17:03 <viveknarasimhan> i presumed the issue you ask here is 1364215...
15:17:17 <Swami> viveknarasimhan: yes
15:17:43 <viveknarasimhan> yes, that is seen in devstack if nodes start much earlier than controlelr
15:18:24 <Swami> viveknarasimhan: have you posted the patch for the SNAT namespace deletion that you added to the SNAT patch
15:18:40 <viveknarasimhan> i posted that here:
15:18:41 <Swami> as a separate patch
15:18:49 <viveknarasimhan> and then it was Abandoned by armax
15:18:57 <viveknarasimhan> since similar patch was posted by Carl here:
15:19:04 <Swami> ok
15:19:40 <Swami> There are couple of other patches that are pending review.
15:19:48 <Swami> SNAT patch that gets rid of hints.
15:19:58 <viveknarasimhan> is there plan to divide teh 'Refactor hints' patch
15:20:02 <viveknarasimhan> into two
15:20:07 <viveknarasimhan> or that is already being pursued?
15:20:21 <Swami> Carl_baldwin: mentioned that he is trying to break it up into chunks instead of pushing it as a single patch
15:20:38 <Swami> So I am waiting on it.
15:20:44 <carl_baldwin> Swami: I posted a link to my proposed work in the patch.
15:20:47 <viveknarasimhan> swami: ok
15:21:06 <Swami> carl_baldwin: Yes I took a look at the proposed patch.
15:21:39 <Swami> carl_baldwin: are you planning to push the individual patches to gerrit
15:22:19 <carl_baldwin> Swami: I wanted your thoughts on it first.  Then, either you or I could post them to gerrit.
15:22:47 <Swami> ok, I will review it once again today and will ping you
15:23:02 <Swami> and then we can proceed
15:23:05 <carl_baldwin> viveknarasimhan: I did not realize that you were duplicating that patch I posted.
15:23:20 <viveknarasimhan> Carl: that is ok carl
15:23:33 <viveknarasimhan> i did that patch as it fixed teh SNAT namespace deletion
15:23:39 <viveknarasimhan> in conjunction with the refactor hints patch
15:23:40 <carl_baldwin> viveknarasimhan: I just happened across it after having already approved the original patch that changed the name.
15:23:49 <viveknarasimhan> so i put that change as dependency to refactor hints patch
15:24:17 <carl_baldwin> viveknarasimhan: ok
15:24:52 <Swami> The migration patch also is waiting for review.
15:24:55 <viveknarasimhan> carl: thanks for finding and pushing that up
15:24:56 <Swami> #link https://review.openstack.org/#/c/105855/
15:25:36 <Swami> There is also another patch that is related to migration that I pushed in yesterday
15:25:40 <Swami> #link https://review.openstack.org/#/c/118491/
15:26:17 <Swami> mrsmith: Did you get a chance to review the migration patch, I cleaned it a bit after your commit.
15:26:28 <mrsmith> not yet
15:26:33 <mrsmith> I will this morning
15:26:41 <Swami> mrsmith: thanks
15:27:19 <Swami> I think we are done with bugs
15:27:28 <Swami> #topic Services
15:27:37 <Rajeev> I have one review targeted for Juno-3 waiting
15:27:42 <Rajeev> Stop tracking connections in DVR FIP Namespace: https://review.openstack.org/#/c/116412/
15:27:46 <Swami> The FWaaS team have posted a patch for DVR N-S support.
15:28:18 <Rajeev> Swami: sorry to interrupt, since juno-3 is tomorrow
15:28:49 <carl_baldwin> Rajeev: Actually, with the current gate pileup, nothing more will make juno-3.
15:28:54 <Swami> Ok thanks
15:29:02 <atiwari> hockeynut, yt?
15:29:25 <hockeynut> atiwari yessir
15:29:32 <Rajeev> carl_baldwin: thanks.
15:29:38 <Swami> carl_baldwin: Does it mean even the bug fixes will not go through Juno 3
15:29:49 <atiwari> I have addressed some of your comments in my latest patch
15:29:59 <carl_baldwin> Bug fixes can still go in after juno-3.
15:30:07 <atiwari> I have also answered your questions
15:30:26 <Swami> carl_baldwin: thanks that what I wanted to confirm
15:30:29 <atiwari> regarding the \ comment, is it ok to address in separate cr?
15:30:36 <hockeynut> certainly
15:30:36 <atiwari> hockeynut, ^
15:30:41 <Swami> atiwari: what patch are you talking about
15:31:11 <atiwari> sorry wrong room, I am very sorry
15:31:20 <Swami> its ok
15:31:22 <carl_baldwin> atiwari: :)
15:31:33 <atiwari> carl_baldwin, :)
15:31:40 <Swami> Let me get back to the Services discussion
15:31:48 <Swami> The FWaaS team have posted a patch for DVR.
15:31:52 <Swami> #link https://review.openstack.org/#/c/113359/
15:33:01 <Swami> FWaaS team had some issues in their rules not getting honored by the SNAT namespace.
15:33:39 <Rajeev> Swami: I have couple of DVR patches I am working on, will try to take a look at FWaaS alongwith
15:34:02 <Swami> SumitNaiksatam requested for a half hour call, if possible we should address that issue, since this patch has to go in today.
15:34:39 <Swami> amotoki: hi
15:34:54 <amotoki> Swami: hi
15:35:44 <Swami> Ok, we can go through the FWaaS issues with the FWaaS team today.
15:36:14 <Swami> #topic horizon
15:36:21 <Swami> amotoki: hi
15:36:38 <amotoki> I just got home a while ago :)
15:37:02 <amotoki> patch series of DVR support in horizon have been merged.
15:37:03 <Swami> amotoki: did you get a chance to work on the router edit page on horizon, sorry just checking.
15:37:38 <amotoki> enable_distributed_router flag is added to horizon settings.
15:37:52 <amotoki> Swami: both patches have been merged.
15:37:53 <Swami> amotoki: great!
15:38:11 <Swami> amotoki: can you provide me the links to both patches
15:38:40 <Swami> If you don't have it handy, it can do a look up in gerrit.
15:39:31 <amotoki> Swami: looking for it but my browser is now slow...
15:39:54 <Swami> amotoki: fine, if you can send me an email later or I can take a look at it.
15:40:03 <Swami> amotok: no worries
15:40:15 <Swami> amotoki: thanks for your help
15:40:32 <amotoki> https://review.openstack.org/#/q/project:openstack/horizon+branch:master+topic:bp/enhance-horizon-for-dvr,n,z
15:41:08 <Swami> amotoki: Thanks for the links
15:41:42 <Swami> amotoki: bhooshan: great job. Well done in a short span
15:41:48 <viveknarasimhan> amotoki: thanks a lot for your timely help in horizon for dvr
15:41:59 <Swami> #topic Open-Discussion
15:42:17 <Swami> Any other open discussion for the DVR
15:42:44 <amotoki> one question. how does DVR experimental job run?
15:43:00 <amotoki> do we use multinode testing?
15:43:08 <Swami> From next week let us plan on addressing some features for the Kilo
15:43:29 <Swami> amotoki: I don't think the multinode is up yet
15:43:38 <carl_baldwin> amotoki: Still no multi-node testing.  I think there has been some slow progress on it but it is not here.
15:43:43 <Swami> we are still using the single node setup.
15:44:13 <Swami> amotoki: there was a patch for review on mutlinode but it is still under review.
15:44:27 <carl_baldwin> I had a question just come up about attaching a vlan network to a distributed router.
15:44:50 <carl_baldwin> Will there be an error if one attempts to do this or will it fail silently under the hood?
15:45:01 <amotoki> Swami: carl_baldwin: thanks. will check the review.
15:45:03 <Swami> viveknarasimhan: any comments on the vlan
15:45:29 <viveknarasimhan> vlan
15:45:41 <viveknarasimhan> is not supported in teh current architecture
15:45:48 <Swami> viveknarasimhan: see the message above from carl_baldwin
15:46:05 <Swami> will it throw an error or fail silently
15:46:18 <viveknarasimhan> i have to check about routing.
15:46:25 <viveknarasimhan> switching will continue to work e
15:46:45 <viveknarasimhan> it will not throw errors
15:46:52 <viveknarasimhan> if network type is vlan, we donot apply any flows
15:46:57 <viveknarasimhan> routing will fail
15:47:03 <viveknarasimhan> switching will continue to work
15:47:21 <carl_baldwin> viveknarasimhan: So, there will be no API error?  The end user will see success from the API?
15:47:39 <viveknarasimhan> yes, he will see success
15:47:57 <viveknarasimhan> not sure if the plugin validates if this is a tunneltype network
15:48:02 <viveknarasimhan> rathern than vlan network
15:48:12 <viveknarasimhan> swami: can you please clarify on that question by carl
15:48:31 <Swami> viveknarasimhan: Will there be an error if one attempts to do this or will it fail silently under the hood?
15:48:31 <viveknarasimhan> it is not checked, so it will succeed
15:48:43 <viveknarasimhan> what do we mean by failure?
15:48:47 <Swami> viveknarasimhan: I think you have answered
15:48:48 <viveknarasimhan> routing will fail
15:48:51 <viveknarasimhan> switching will succeed
15:49:06 <carl_baldwin> viveknarasimhan:  I mean will the API call fail.  Sounds like it will not.
15:49:08 <Swami> from an api perspective it will not through any errors.
15:49:15 <viveknarasimhan> the plugin does not stop adding vlan subnet to dvr
15:49:19 <carl_baldwin> Okay.  Thanks.
15:49:24 <carl_baldwin> I’ll put a bug in to the backlog.
15:50:14 <Swami> carl_baldwin: thanks that would help
15:50:25 <Swami> thanks everyone for joining the call.
15:50:29 <Swami> See you all next week
15:50:38 <Swami> bye
15:50:49 <Swami> #endmeeting