15:18:07 <tmorin> #startmeeting bgpvpn
15:18:09 <openstack> Meeting started Tue Sep  1 15:18:07 2015 UTC and is due to finish in 60 minutes.  The chair is tmorin. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:18:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:18:13 <openstack> The meeting name has been set to 'bgpvpn'
15:18:24 <pcarver> hi
15:18:30 <tmorin> hi paul
15:18:40 <tmorin> #topic announces
15:18:45 <tmorin> hi doude
15:18:58 <matrohon> doude is not available today
15:19:02 <tmorin> ok
15:19:58 <tmorin> we've merged some work since last week
15:20:11 <tmorin> some uninteresting stuff to follow oslo changes
15:20:23 <tmorin> some devstack fixes
15:20:34 <tmorin> but also more significant stuff
15:20:58 <tmorin> the patch for the DB to accept the route_distinguishers field from svinota has been merged
15:21:31 <tmorin> and also the patch to allow us to specify the service_providers for BGPVPN in our own file (/etc/neutron/networking_bgpvpn.conf)
15:21:57 <tmorin> that last one was required since reading service providers from neutron.conf is now deprecated
15:22:43 <tmorin> also a patch to cleanup the bgpvpn_db file and not add checks/filters additionally to what is supposed to be already done through policy.json
15:22:53 <tmorin> #topic work in progress
15:22:56 <tmorin> #undo
15:22:57 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0xa56fb90>
15:23:05 <tmorin> anything else to announce ?
15:23:57 <tmorin> #topic work in progress
15:24:39 <matrohon> the new API is up for review
15:25:06 <matrohon> I just rebased it : https://review.openstack.org/#/c/207371/
15:25:20 <matrohon> I took into account pcarver and tmorin reviews
15:25:29 <tmorin> (it seems it still needs a rebase though)
15:25:52 <tmorin> the neutronclient part is not yet done, I've just started working on it today
15:25:59 <matrohon> It's because svinota's patch just merged
15:26:13 <tmorin> indeed
15:26:37 <tmorin> I think the new API/DB patches mostly look good
15:26:51 <tmorin> we still need to figure out the policy.json part though
15:26:53 <matrohon> I know that doude had some concern about it
15:27:06 <tmorin> even for the old API, it seems that it doesn't do yet what it is supposed to
15:28:03 <tmorin> matrohon: doude comments were mostly minor I would say (some unneeded stuff we could remove)
15:28:07 <matrohon> he feels strange that the "get_bgpvpn" returns a network list, while 'network ' is not a bgpvpn attribute
15:28:43 <matrohon> tmorin : what comments are you talking about
15:28:46 <tmorin> but right now statement like "    update_bgpvpn_connection:export_targets": "rule:admin_only"  " do not work, a random tenant user can update route-targets
15:28:53 <tmorin> at least on my devstack
15:29:11 <tmorin> matrohon: comments he made privately to me
15:30:05 <matrohon> tmorin : can you open a bug concerning the policy issue
15:30:14 <tmorin> yes, I will
15:30:27 <tmorin> need to first check that the issue is not between the chair and the keyboard ;)
15:30:52 <tmorin> I suggested doude to file a bug on the improvements to do
15:31:57 <tmorin> matrohon: I did not get what you meant by "he feels strange that the "get_bgpvpn" returns a network list, while 'network ' is not a bgpvpn attribute"
15:32:10 <pcarver> Is there a local.conf somewhere that I should refer to? I've been reviewing code but haven't actually tested anything locally yet.
15:33:10 <tmorin> the README.md and README-bagpipe.md have the basic info on what you need to add in a local.conf
15:33:29 <matrohon> pcarver : I had some local.conf available on my github : https://github.com/mathieu-rohon/devstack-conf
15:33:32 <pcarver> tmorin: thanks, I'll take a closer look and try to set it up
15:33:34 <tmorin> and matrohon has a devstack with an example local.conf
15:33:35 <tmorin> yes
15:33:46 <pcarver> matrohon: thanks
15:34:26 <matrohon> I didn't try them since few days, maybe they are not compatible with recent patches
15:34:30 <tmorin> matrohon: your local.conf is not updated to add the service provider in the right file
15:34:38 <matrohon> tmorin : +1
15:34:54 <matrohon> I'll update them ASAP
15:35:36 <matrohon> concerning doude's comments : he should review the spec to make them public
15:36:09 <tmorin> matrohon: +1  (or I'll file a bug report...)
15:36:31 <matrohon> but in this patch : https://review.openstack.org/#/c/217038/3/networking_bgpvpn/neutron/db/bgpvpn_db.py
15:36:48 <matrohon> the get_bgpvpn will return the associated networks
15:37:25 <matrohon> and since networks are not an attribute of the bgpvpn resource anymore, he doesn't feel comfortable with this change
15:38:47 <tmorin> matrohon: I understand your/his comment now, but I've honestly no idea whether it should be considered an issue or not
15:40:06 <matrohon> me neither, let's wait for his comment on the spec
15:40:12 <tmorin> yes
15:40:43 <tmorin> the behavior is consistent with what we have written in the specs, and nobody commented that it would be an issue
15:41:27 <tmorin> there is something related that we will have to take care of
15:41:37 <tmorin> GET /bgpvpn/bgpvpns?network={network_uuid}
15:42:09 <tmorin> since networks is not an attribute of BGPVPN, it is a particular case that we'll need to look at
15:43:13 <tmorin> ok, next topic ?
15:43:27 <tmorin> what do we have to discuss ?
15:43:38 <matrohon> #topic driver implementation
15:44:09 <tmorin> I've a pending patch stop having the bagpipe driver prevent the deletion of a network
15:44:12 <tmorin> reviews welcome
15:44:19 <tmorin> #link https://review.openstack.org/218262
15:44:37 <matrohon> I started the patch to change the bagpipe driver so that it take account the split of the API
15:44:50 <tmorin> I have another one cooking so that the bagpipe driver does not need its own mech_driver to be notified about ports coming up or down
15:45:06 <tmorin> not ready and not yet pushed to gerrit though
15:45:33 <tmorin> doude has been making some progress on the contrail driver
15:45:45 <tmorin> still not ready for review either though
15:46:11 <matrohon> concerning 218268
15:46:30 <matrohon> it conflicting with the db change I pushed
15:46:52 <tmorin> 218268 is a nova change ?  typo ?
15:47:16 <matrohon> oups 218262
15:47:28 <matrohon> the one you just mentioned
15:47:40 <matrohon> it is conflicting with https://review.openstack.org/#/c/217038/3/networking_bgpvpn/neutron/db/bgpvpn_db.py
15:48:08 <matrohon> by using a foreign key, I prenvent the deletion of the netwok at the db layer
15:49:28 <tmorin> yes, but during API discussing, it was explained (can't remember by whom) that foreign keys accross openstack project should be avoided
15:50:39 <tmorin> and we also considered that from an API user point of view, there is no reason to block the deletion of the network, we should simply remove associations of this network
15:51:42 <matrohon> let's discuss that during reviews
15:52:09 <matrohon> janscheurich, any update for the ODL driver?
15:52:25 <tmorin> matrohon: ok
15:53:41 <janscheurich> matrohon: unfortunately work has not yet started
15:53:54 <matrohon> janscheurich, ok
15:54:12 <janscheurich> There is a dependency to ODL Neutron Northbound
15:54:39 <janscheurich> BGPVPN must be modelled in Yang
15:54:57 <janscheurich> That needs to be done before the driver can be implemented
15:55:08 <matrohon> janscheurich, fine
15:55:20 <matrohon> #topic : tokyo summit
15:55:43 <matrohon> unfortunately, the bgpvpn talk has not been selected
15:56:25 <matrohon> svinota, pcarver, janscheurich : will you attend the next summit?
15:56:29 <svinota> but we still can discuss it, being there
15:56:36 <pcarver> matrohon: I'll be there
15:56:40 <svinota> matrohon, yep, I will be at the summit
15:56:57 <matrohon> tmorin and I will be there too
15:57:07 <janscheurich> We are still in planning phase. But someone from Ericsson Aachen will likely be there
15:57:12 <tmorin> I'll be in Tokyo too
15:57:20 <janscheurich> Either me or Tim
15:57:26 <pcarver> I would like to schedule an ad hoc session
15:57:38 <tmorin> pcarver: +1
15:58:06 <pcarver> We've been talking to lots of people about BGPVPNs in various contexts and Margaret and I would like to try to get everyone together in the same room at some point
15:58:24 <matrohon> pcarver : makes sense
15:58:30 <janscheurich> pcarver +1
15:58:31 <tmorin> a lot of sense, yes!
15:59:11 <matrohon> pcarver : do you think about a dedicated design summit session?
15:59:22 <pcarver> I'm expecting an Etherpad of design session proposals to come out soon. I was going to add BGPVPN to it
15:59:39 <matrohon> pcarver : +1
15:59:43 <tmorin> +1
15:59:51 <janscheurich> tmorin: What about the BGPVPN TechTalk you suggested?
15:59:55 <pcarver> If it doesn't get selected then I figured just an email to the -dev list the week of the summit to announce an ad hoc meetup spot
16:00:14 <matrohon> pcarver : thanks
16:00:30 <matrohon> ok we have to leave!
16:00:41 <tmorin> thanks everyone
16:00:45 <matrohon> #endmeeting
16:00:45 <pcarver> bye
16:00:48 <tmorin> bye bye
16:00:49 <janscheurich> Thanks, bye
16:00:50 <svinota> thanks
16:01:00 <openstack> adrian_otto: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
16:01:12 <matrohon> oh, im'not chairing this meeting :)
16:01:20 <matrohon> #endmeeting
16:01:23 <adrian_otto> who started the previous meeting, matrohon?
16:01:30 <matrohon> tmorin ^
16:01:34 <mfalatic> o/
16:01:43 <adrian_otto> mfalatic: hold on just a moment
16:01:44 <matrohon> tmorin, can you end the meeting?
16:02:27 <tmorin> #endmeeting