15:05:35 #startmeeting bgpvpn 15:05:37 Meeting started Tue Aug 4 15:05:35 2015 UTC and is due to finish in 60 minutes. The chair is matrohon. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:05:38 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:05:40 The meeting name has been set to 'bgpvpn' 15:06:34 ok so no bing annoucement for me today, except taht I'm going to chair the meeting since tmorin is in vacation for 3 weeks 15:07:44 we decided to keep the spec under the neutron-spec repo untill we have our first release 15:08:39 we will move the spec to the bgpvpn repo and we will update it according to what is shipped with the first release 15:08:40 hi all 15:08:46 hi svinota 15:09:24 pcarver, svinota : any announcement ? 15:09:36 nope. 15:10:09 nothing from me. I think I've reviewed everything I'm aware of. 15:10:51 #info : the implementation of the opencontrail driver has begun : https://review.openstack.org/#/c/202806/ 15:11:25 #undo 15:11:25 Removing item from minutes: 15:11:46 #topic announcement 15:11:58 #info : the implementation of the opencontrail driver has begun : https://review.openstack.org/#/c/202806/ 15:12:40 #info : tmorin has started to introduce a policy.json for bgpvpn : https://review.openstack.org/#/c/207914/ 15:13:38 #info : matrohon is working on the API split : https://review.openstack.org/#/c/207371/ 15:14:34 #info : svinota is working od RD attribute addition : https://review.openstack.org/#/c/206536/ 15:14:57 can we discuss this attribute here svinota? 15:15:09 matrohon, yep 15:15:23 #topic RD Attribute 15:16:09 so as I reviewed, I'm concerned since not every driver will support this attribute 15:16:30 yep. And thus the driver can ignore it 15:17:24 this gone lead to an unexpected behavior if the user specifies a RD : the driver won't use it but the end user expect the driver to use it 15:18:06 so we have a simple choice 15:18:31 either to raise an exception, or to have a capabilities discovery + exception 15:19:33 the capabilities discovery seems cleaner and such a framework already exists : it is the extension framework 15:19:34 the exception is required anyway, since the user (or UI) may ignore the discovery, if I understand this correctly 15:20:32 I think the exception framework will automatically generate an exception if the user want to use an attribute extension that is not supported by the implementation 15:20:46 matrohon, so then I'm to implement the discovery 15:21:00 should not be too hard, I hope 15:21:09 cool 15:21:48 I don't think it's very hard, we can look at that together 15:22:30 it's just an extension of the bgpvpnconnection object 15:22:41 ok, thanks 15:23:10 I will try to do it and submit to review. 15:23:35 #action : svinoto to work on an extension concerning the RD attribute 15:23:43 #undo 15:23:44 Removing item from minutes: 15:23:48 #action : svinota to work on an extension concerning the RD attribute 15:25:20 this kind of attribute extension may also be used to scan the "technique" parameter 15:26:02 not sure if it's a good idea, but this option is also backend dependent 15:27:13 we'll see once the RD attribute extension will exist 15:27:46 we gone have a better idea of what is possible with the attribute extensions framework 15:28:24 #topic Tests 15:28:48 I really want to improve the test coverage of the bgpvpn code 15:29:12 this will help debugging and CI 15:29:59 I'm trying to learn from existing project (vpnaas, etc...) but the unit test framework in neutron is quite tricky 15:30:32 #action : matrohon to work on test coverage improvment 15:31:26 #topic open discussion 15:31:45 anybody wants to discuss something about bgpvpn 15:31:50 ? 15:32:52 ok, pcarver, svinota, thanks for attending the meeting 15:33:07 #endmeeting