15:05:54 <tmorin1> #startmeeting bgpvpn 15:05:55 <openstack> Meeting started Tue Sep 26 15:05:54 2017 UTC and is due to finish in 60 minutes. The chair is tmorin1. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:05:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:05:58 <tmorin1> hi everyone 15:05:59 <openstack> The meeting name has been set to 'bgpvpn' 15:06:06 <tmorin1> I'm not sure who is around 15:06:16 <tmorin1> pcarver ? doude ? 15:06:29 <doude> hi 15:06:41 <tmorin1> I was going to propose a very short meeting because I'm in a face2face meeting right now 15:06:43 <tmorin1> hi doude 15:07:27 <tmorin1> doude: on the route-control first implementation change (https://review.openstack.org/#/c/481125/) 15:07:35 <tmorin1> I've seen your comments (thanks!) 15:07:52 <tmorin1> I'll address them soon (this week I hope) 15:08:05 <tmorin1> the comments on class inheritance seem reasonable 15:08:33 <tmorin1> I think I also agree with you that the check for port_association operation are not needed 15:08:53 <doude> ok I just think it but not tried, need to be validated 15:08:55 <tmorin1> the other comment abour _validate_port : I don't know, but I'll have a look 15:09:05 <pcarver> hi, sorry, I'm hear but still on a voice conf call that's running long. 15:10:57 <doude> I think you should also validate port on update, no? 15:11:34 <tmorin1> pcarver: hi, we're doing only a short meeting / please comment on whatever you want 15:12:11 <tmorin1> doude: the only thing that validate port checks is that the port exists => since port_id is read-only, no need to check on update again 15:12:44 <tmorin1> we can group everything in one method, but I don't think it really is a huge improvement 15:12:58 <tmorin1> doude: ^, do you think this matters a lot ? 15:13:17 <doude> ha ok I missed that the port_id cannot be update in a port assoc 15:13:24 <doude> ok not it's correct 15:17:42 <tmorin1> doude: ok 15:17:59 <tmorin1> I don't have anything precise to discuss today 15:18:06 <tmorin1> ah yes, there is one thing 15:18:24 <tmorin1> doude: I updated https://review.openstack.org/#/c/499943/ based on your comments 15:18:27 <tmorin1> I think it's ready 15:19:07 <tmorin1> but we need one thing : checking that the change of update pre commit dict does not impact OVS driver (the only driver currently using precommit) 15:19:29 <tmorin1> so doude, if you can review the new PS that will help 15:19:51 <doude> ok 15:20:28 <doude> about that tmorin1, I saw in the ML2 plugin a context manager mechanism. I don't know if that can fit here 15:20:57 <doude> but I know ml2 have a similar mechanism with pre and post hook to mechanism driver 15:21:13 <doude> perhaps there is some code we can use here 15:21:27 <doude> or perhaps it's overkill for our need 15:21:32 <doude> s 15:22:03 <tmorin1> doude: would you have link ? 15:22:36 <tmorin1> (I know that we need to use the more recent DB facade in BGPVPN, but its probably not what you have in mind here) 15:22:43 * doude tries to find it again 15:23:08 <doude> ha yes I also see that new DB facade 15:25:10 <doude> tmorin1 something like that https://github.com/openstack/neutron/blob/master/neutron/plugins/ml2/managers.py#L409 15:25:47 <doude> if I understand correctly, that permits to rollback in case an error appear during the hook call to MD 15:26:57 <tmorin1> doude: I think this method is more about how to behave when calling multiple drivers and one fails, isn't it ? 15:27:16 <tmorin1> doude: the ML2 DB code is not (I think) in the MD manager 15:27:33 <doude> ha ok my bad 15:27:47 <tmorin1> I think the pattern for pre/post commit hooks wrt to DB calls was borrowed from ML2 by matrohon at the time 15:28:00 <tmorin1> so I expect that we already have code similar to ML2 15:34:26 <doude> tmorin1: there is a context per resource mechanise, ie. the port case https://github.com/openstack/neutron/blob/master/neutron/plugins/ml2/plugin.py#L561 15:34:50 <doude> but not sure is necessary in our case, we don't have to mange multiple driver 15:41:32 <tmorin1> ok 15:41:33 <tmorin1> sorry 15:41:36 <tmorin1> I was interrupted 15:41:47 <tmorin1> I'm not sure we have thing important to discuss this week 15:42:01 <tmorin1> there are a few things in the pipe, but the discussion is happening via gerrit in most cases 15:42:07 <tmorin1> I propose to close the meeting 15:42:12 <tmorin1> thanks doude! 15:42:23 <tmorin1> #endmeeting