15:29:00 <tmorin> #startmeeting bgpvpn
15:29:01 <openstack> Meeting started Tue Jul 11 15:29:00 2017 UTC and is due to finish in 60 minutes.  The chair is tmorin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:29:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:29:05 <openstack> The meeting name has been set to 'bgpvpn'
15:30:11 <tmorin> #topic rollcall
15:30:16 <tmorin> let's see who's around today
15:30:58 <pcarver> hi
15:31:24 <tmorin> hi paul
15:31:46 <tmorin> maybe doude ? bobmel ?
15:31:56 <tmorin> I know matrohon is on holidays
15:32:09 <doude> hi
15:34:37 <tmorin> hey doude
15:34:41 <tmorin> ok let's start
15:34:49 <tmorin> #topic agenda
15:34:55 <tmorin> here is what I see for today
15:35:07 <tmorin> a) bgpvpn-routes-control extension work in progress
15:35:15 <tmorin> b) driver feature support matrix
15:35:23 <tmorin> c) bug review
15:35:33 <tmorin> anything else ?
15:35:55 <tmorin> ah yes: d) new tempest tests
15:36:39 <tmorin> #topic bgpvpn-routes-control extension work in progress
15:37:05 <tmorin> while implementing the framework and DB for the bgpvpn-routes-control extension, I found out about issuess in the API definition in neutron-lib
15:37:27 <tmorin> https://review.openstack.org/#/c/481088/ is in flight to fix these, mostly landed now
15:38:06 <tmorin> the change with the framework/db work for port associations is here: https://review.openstack.org/#/c/481125/
15:38:12 <tmorin> I think it's fairly complete
15:38:37 <tmorin> not passing tempest tests because we'll need a neutron-lib release with our API extension first
15:39:12 <tmorin> reviews are welcome, in particular on the driver hooks (doude, if you want to have a look)
15:40:13 <tmorin> that said, even with this work, given that I'll be off until end of July starting next Friday, I don't plan to try to land the bgpvpn-routes-control extension for Pike
15:40:27 <tmorin> comments ?
15:40:48 <doude> tmorin: ok
15:41:12 <tmorin> thanks!
15:42:55 <tmorin> #topic driver feature support matrix
15:43:17 <tmorin> pcarver has pushed a change to add a driver feature matrix to the doc:
15:43:40 <tmorin> thanks pcarver !
15:43:44 <tmorin> #link https://review.openstack.org/#/c/482357/
15:43:56 <tmorin> discussion ongoing on the content
15:44:05 <tmorin> it will definitly be a good addition
15:44:25 <tmorin> this also was the trigger to identifiy that we need updated doc on opencontrail and odl drivers
15:44:46 <tmorin> to document that their v1 drivers are obsolete, and provide minimal proper info on how to use the non-obsolete ones
15:45:00 <pcarver> I'll provide an update. I didn't actually realize that there was a new opencontrail driver
15:45:27 <tmorin> it's in opencontrail repo
15:45:28 <tmorin> doude: ok to push a change to update the doc for opencontrail ?
15:46:54 <doude> yes tmorin. But i also think to push a patch to completely remove the old driver
15:47:45 <doude> I think is better to remove it now as nobody use it for the moment and only let the active one
15:48:13 <pcarver> I agree we should remove the old one but maybe there should be a deprecation notice first, then a removal as a second step
15:48:21 <tmorin> doude: it depends whether or not you want to provide a driver for contrail < 4
15:48:35 <tmorin> doude: wouldn't this be a reason to /not/ remove it yet ?
15:48:55 <tmorin> doude: perhaps instead have it raise a runtime warning that it is obsolete
15:49:28 <doude> the v1 code was more a PoC than a release, it does not use the correct way to se RT on VRF
15:50:28 <doude> we can deprecate it if you prefer
15:50:42 <doude> what will be the deprecation schedule?
15:52:33 <tmorin> doude: I think it has to be determined consistently with the deprecation schedule for the last version of opencontrail that does not have what is required by the v2 driver  (3.2 ?)
15:52:33 <pcarver> I guess we don't know for sure who's using anything, but I suspect you're right that nobody is really using it in production so a short time should be fine.
15:53:07 <doude> pcarver: ok
15:53:09 <pcarver> I'm just thinking maybe update with a pointer to the replacement and then a month or two later remove it completely.
15:53:52 <doude> tmorin: no it's not recommended to use the v1 with contrail release < 4 as there is no migration path to move to the new driver
15:54:26 <tmorin> doude: then a "short" deprecation schedule would make sene
15:54:27 <tmorin> sense
15:54:36 <doude> pcarver: I don't think we can work with month step but with release step
15:54:57 <tmorin> pcarver: yes
15:55:15 <tmorin> "will be removed in Queens" would be reasonable I think
15:55:31 <pcarver> tmorin: sure, that's fine
15:55:35 <doude> ok
15:55:38 <tmorin> ok, next topic ?
15:55:48 <tmorin> #topic bug review
15:55:50 <tmorin> the
15:56:16 <tmorin> there were/are a few things needed for the horizon dashboard, to adapt to recent horizon changes
15:56:24 <tmorin> some fixes have merged (cedric's work)
15:56:32 <tmorin> and some others are in flights
15:57:12 <tmorin> #topic new tempest tests
15:57:21 <tmorin> some new tempest tests are beeing refined
15:57:32 <tmorin> I think they should soon be ready for merging
15:58:07 <tmorin> #link https://review.openstack.org/#/c/480142/
15:58:42 <tmorin> #topic open discuss
15:58:45 <tmorin> anything else ?
15:58:49 <tmorin> 2min left :)
15:58:54 <tmorin> doude ? pcarver ?
15:59:05 <doude> ok for me
15:59:08 <pcarver> nothing else for me
15:59:30 <tmorin> ah yes, one more thing: no meeting next week
15:59:48 <tmorin> two weeks for now: I'm not sure (I'll be off, matrohon may be back)
15:59:54 <tmorin> three weeks from now: ditto
16:00:02 <tmorin> thanks folks!
16:00:20 <tmorin> leaving the floor for next meeting...
16:00:21 <tmorin> bye
16:00:26 <tmorin> #endmeeting