16:05:28 <rkukura> #startmeeting networking_ml2 16:05:28 <openstack> Meeting started Wed Jun 14 16:05:28 2017 UTC and is due to finish in 60 minutes. The chair is rkukura. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:05:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:05:31 <openstack> The meeting name has been set to 'networking_ml2' 16:05:40 <rkukura> #topic Agenda 16:06:04 <rkukura> #link https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_June_14.2C_2017 16:06:10 <rkukura> Same agenda as recent meetings 16:06:15 <rkukura> Anything to add? 16:06:34 <rkukura> #topic Announcements 16:06:39 <rkukura> I have no announcements 16:06:42 <rkukura> Anyone else? 16:07:02 <trevormc> nothing too much, looks like one of my patches for qinq is about to land 16:07:14 <rkukura> I missed last week, and don’t see minutes, but Sukhdev said there was a meeting 16:07:23 <rkukura> I’ll also most likely miss next week 16:07:28 <reedip_> o/ 16:07:43 <Sukhdev> Sorry for late entry 16:07:50 <trevormc> hi reedip_ 16:08:09 <rkukura> hi Sukhdev, reedip_ 16:08:22 <Sukhdev> I must have mistyped the meeting name. Will look for the minutes and pist 16:08:25 <rkukura> #topic RFEs/bugs 16:08:41 <rkukura> trevormc: sounds like you may have some updates for us 16:09:08 <trevormc> so the refactor patch is about to land 16:09:09 <trevormc> https://review.openstack.org/#/c/458531/ 16:09:22 <trevormc> but I had a question on my other patch that I wasn't sure on how to answer 16:09:43 <trevormc> https://review.openstack.org/#/c/446047/34//COMMIT_MSG@23 16:11:01 <trevormc> if I could have someone's input on yamamoto's comment, that would be appreciated. I thought the driver would serve as a general purpose way for using QinQ networks with any backend, not just ovs. 16:12:11 <trevormc> i didnt have any other updates 16:12:35 <rkukura> I’m not sure I can help with yamamoto’s comment 16:12:58 <trevormc> that's ok. 16:13:16 <rkukura> seems like following up in gerrit makes most sense, or starting an email thread 16:13:29 <sadasu> my simple answer would be that each ml2 mech driver handled this case their own way 16:13:31 <trevormc> yeah agreed 16:13:43 <trevormc> sadasu: that's what I thought too 16:14:20 <sadasu> that meant that some config was specified to a specific mech driver to handle it within its own logic 16:14:20 <trevormc> Sukhdev_, on another note, I reached out to rosella, she commented on my bug for sriov vlan filter. She says it possible and the proposed way in the bug is the way to go. 16:14:40 <sadasu> but, now you are providing a generic ml2 way of handling this 16:14:41 <Sukhdev_> trevormc : cool 16:15:04 <sadasu> I had to support this functionality in my ucsm plugin 16:15:31 <trevormc> sadasu, oh that's interesting, how did it work? 16:15:47 <sadasu> and due to various reasons did not go the route of proposing a generic ml2 solution 16:16:18 <trevormc> ok maybe an ML discussion would be appropriate, about how to handle QinQ networks. 16:16:45 <sadasu> the "internal VLANs" , borrowing from yamamoto's terminology 16:17:27 <Sukhdev_> trevormc: I just read yamamoto's comment - he is basically saying that this should be handled in the back-end by the plugins/drivers - instead of exposing it into the API 16:17:30 <sadasu> was specified as a list to the mech driver and then the outer-Vlan or the VM - VLAN was managed via neutron / ml2 16:18:26 <Sukhdev_> sadasu trevormc : yamamoto is essentially saying that you solve it the way sadasu did 16:19:03 <trevormc> Hmm interesting, so what should we do moving forward? I guess I could WF -1 the refactor patch 16:19:12 <sadasu> Sukhdev_, trevormc: agreed and it is a valid question 16:19:42 <trevormc> it was about to merge but jenkins hasn't passed yet, (luckily ? 16:19:43 <Sukhdev_> trevormc : so, you can answer him in two ways : 16:20:00 <sadasu> trevormc: just explain that it is a generic way and each mech driver should not have to re-invent 16:20:11 <Sukhdev_> 1) you are solving it in a more generic manner so everybody can use it 16:20:27 <Sukhdev_> 2) or you will follow the same trick which sadasu did 16:20:37 <Sukhdev_> but, since, this is ready to merge, I will stay with it 16:20:52 <sadasu> I agree with Sukhdev_ 16:20:53 <trevormc> i was going to say that but.. i may be a little bias 16:20:54 <trevormc> haha 16:21:19 <trevormc> ok, I'll let him know and make a comment in review about it 16:21:51 <trevormc> - nothing else from, me thanks for the support! 16:21:57 <Sukhdev_> trevormc : If you send me the link, I will be happy to add the comment for you :-) 16:22:04 <rkukura> thanks trevormc! 16:22:20 <trevormc> #link https://review.openstack.org/#/c/446047/ 16:22:35 <rkukura> OK, any other RFEs/bugs to discuss today? 16:23:05 <rkukura> #topic Open Discussion 16:23:09 <rkukura> Anythign to discuss today? 16:23:35 <rkukura> If not, we can wrap up early 16:24:04 <rkukura> Thanks everyone! 16:24:11 <trevormc> o/ 16:24:14 <sadasu> thanks! 16:24:16 <rkukura> #endmeeting