16:08:06 #startmeeting networking_ml2 16:08:07 Meeting started Wed Mar 15 16:08:06 2017 UTC and is due to finish in 60 minutes. The chair is Sukhdev_. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:08:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:08:11 The meeting name has been set to 'networking_ml2' 16:08:29 #topic: Agenda 16:08:29 #link: https://wiki.openstack.org/wiki/Meetings/ML2#Meeting_March_15.2C_2017 16:08:37 trevormc : yes 16:08:55 ok I will push it up in a moment 16:09:08 #topic: Anouncements 16:09:38 I do not have any specific announcements - I missed neutron core meeting this week 16:09:55 anybody has any announcements for the team? 16:10:19 #topic: Enhanced Agenda Items 16:10:38 We are adding two new items for discussion to the agenda 16:11:00 1) QinQ type driver 16:11:28 2) ML2 Driver extension enforcement 16:11:57 hey sorry my browser froze 16:12:30 rkukura : can you go first? while trevormc_ fixes the browser issue 16:12:56 actually, if I can have another couple minutes, I’ll post my RFE 16:13:05 it may be 16:13:21 rkukura : sounds good - we are here 16:14:00 trevormc_ : let me know when you are ready 16:14:41 Sukhdev_: ok, browser fixed now. 16:15:13 trevormc__ : cool - so, you go first while rkukura is wrapping up his RFE 16:15:33 https://review.openstack.org/446047 16:16:36 I know there are some things wrong with this patch. I'm here to volunteer my time to develop and debug. 16:17:14 I'm hoping neutron cores will push me in the right direction so we can land the new network type driver. 16:17:59 trevormc__ : makes sense 16:18:09 #link: https://review.openstack.org/446047 16:18:44 trevormc__ : I will review this patch - one quick observation 16:19:11 I noticed you are adding new DB table. You may want to add the migration scripts as well 16:19:39 Please do! Or I can learn and do it myself. 16:20:00 There are a couple alternatives to this approach. I guess I shouldve put it in the commit message 16:21:04 trevormc__ : it is very trivial step that needs to be added - but, we can worry about it later, lets first make sure this driver does the right things 16:21:45 Instead of adding another table there was talk about modifying the segmentation_id field to hold the values in a dictionary. I think the Trunk team wanted to use this too. 16:22:11 trevormc__ : I like that approach as well - 16:22:33 I found it difficult to use the SegmentTypeDriver with that approach 16:22:45 because it relies on the allocated field. 16:22:45 there was a proposal a while back to generalize the “provider” attributes that may be worth digging up, but a dict might suffice 16:22:50 the idea of segmentation_type and segmentation_id is to be able to accomodate various types 16:23:47 rkukura : +1 16:24:50 trevormc__ : you have RFE/Bug filed for this feature, right? 16:25:13 Sukhdev, no RFE/bug is filed for specifically this. 16:25:29 I thought I could use https://bugs.launchpad.net/neutron/+bug/1604222 16:25:29 Launchpad bug 1604222 in neutron "[RFE] Implement vlan transparent for openvswitch ML2 driver" [Wishlist,Confirmed] - Assigned to Bo Chi (bochi-michael) 16:25:57 because in the reviews sean commented he prefers a network type driver to solve this problem. 16:26:13 https://review.openstack.org/#/c/342755/ 16:28:19 trevormc : I see - 16:29:07 trevormc : So, if you want to use the same bug for this, then you should refer to it in your patch - so, that people can tie these together and get the history 16:29:26 Sukhdev_: ok I was just thinking about that 16:30:00 trevormc : right - update the commit message to include the appropriate links 16:32:31 this week is very tight for me. I will try to review the patch and post comments either over the weekend or early next week - 16:32:55 I am sure rkukura will have good feedback for you as well 16:33:18 Thank you, I will be working on the unit tests in the meantime. 16:33:21 I’ll try - added myself as a reviewer 16:33:33 anything else on this type driver? 16:33:48 no thats all, thanks. 16:34:34 #topic: Extensions enforcement discussion 16:34:42 Ok 16:34:47 rkukura : are you ready? 16:35:04 At the PTG, I had promised to file an RFE for this 16:35:19 I finally got to it this morning 16:35:36 excellent 16:35:41 #link https://bugs.launchpad.net/neutron/+bug/1673142 16:35:41 Launchpad bug 1673142 in neutron "[RFE][ML2] Enforce extension semantics" [Undecided,New] 16:36:03 I tried to just capture what had been discussed at the PTG 16:36:52 I think the only substantive change was realizing that the enforcement needed to be done entirely within the transacation that commits a port binding, otherwise extension values might change before the binding is committed 16:37:42 The RFE is a bit dense, so I don’t think it makes sense to review it in detail here now, but maybe we could discuss it next week 16:38:32 rkukura : right, I tried to scan through it - need time to review 16:39:11 I’ll post a link on the email thread that kevinbenton started on this 16:39:29 rkukura : +1 16:39:59 that’s it for me 16:40:19 trevormc : you may want to review this as well - this may apply to the RFE that you shared with us at PTG 16:40:40 thanks rkukura 16:40:45 Sukhdev, I bookmarked it 16:41:03 will do 16:41:12 Anything else? 16:42:14 not from me 16:42:17 not from me either 16:42:22 since we do not have dasanind here, I am going to skip the port binding stuff - unless anybody has anything to discuss about it 16:42:54 #topic: Open Discussion 16:43:18 The Boston summit is fast approaching - 16:44:09 rkukura : you are going to be our official guide :-):-) 16:44:22 I’ll miss Monday, though 16:44:35 or shall I say I nominate you for the job :-) 16:45:11 snow should be gone by then 16:45:45 we had 78 degrees yesterday - when you were dealing with the snow :-):-) 16:46:15 I was in FL for the last storm, so it was nice to be here this time 16:46:51 of course it was only in the 50s and 60s last week when I was in SJC 16:47:04 as long as you are having fun - that is what matters... 16:47:14 right 16:47:29 I can stand snow for a day or so... not more than that 16:47:30 lets wrap up so I can get some lunch ;) 16:47:34 I am chicken :-):-) 16:47:35 +1 16:47:50 OK - 16:48:03 Thanks for discussion folks 16:48:07 bye 16:48:10 \o 16:48:12 thanks Sukhdev_! 16:48:13 #endmeeting