14:00:08 #startmeeting neutron_l3 14:00:08 Meeting started Wed Jun 30 14:00:08 2021 UTC and is due to finish in 60 minutes. The chair is liuyulong. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:08 The meeting name has been set to 'neutron_l3' 14:00:16 slaweq: I included neutron-core to tap-as-a-service-core for now (from gerrit web GUI only) 14:01:14 Hi there, good evening. 14:02:03 liuyulong: Hi 14:02:40 I have no announcements today, so we can run to bugs section directly. 14:02:44 #topic Bugs 14:03:12 #link http://lists.openstack.org/pipermail/openstack-discuss/2021-June/023362.html 14:03:37 #link http://lists.openstack.org/pipermail/openstack-discuss/2021-June/023190.html 14:03:53 hi 14:04:39 The past two weeks was relatively quiet, not have too many bugs 14:05:40 #link https://bugs.launchpad.net/neutron/+bug/1933092 14:06:42 This bug seems to be related to 2 old ones: 14:07:24 #link https://bugs.launchpad.net/neutron/+bug/1774459 14:09:04 and https://bugs.launchpad.net/neutron/+bug/1916761 14:09:35 #link https://review.opendev.org/c/openstack/neutron/+/779464 14:09:56 but the 16.3.2 has the revert patch of the DVR ARP contorl plane. 14:13:25 I will try to use the scripts from the reportor to reproduce the issue. 14:14:26 Next one 14:14:27 #link https://bugs.launchpad.net/neutron/+bug/1933273 14:15:23 This is clearly a bug of code which does not handle the MTU change action well. 14:16:24 The code use the wrong namespace for the device name of qr-e05ab302-1b, there is no qf-dev in snat namespace. 14:16:40 sorry I have to leave now, bye 14:17:47 regarding to the device mtu, the sg-device in the snat namespace should be handled to change the MTU as well, because it is one port from the user's network. 14:18:36 Next 14:18:38 #link https://bugs.launchpad.net/neutron/+bug/1933502 14:19:30 This one looks like an RFE which is trying to change current L3 floating IPs processing procedure. 14:19:41 For now, the steps are: 14:20:23 1. floating IP updated, 2. notify L3 agent that this floating IPs router is updated. 3. L3 agent sync the router info. 4. L3 agent reprocess all router related resources 14:20:46 Yes, it is definitely heavy. 14:21:49 If resource cache related RPC can be used, this can be changed to: 14:22:42 1. floating IP updated. 2. OVO object update event send out. 3. L3 agents which are interested in this event can do the following work. 14:23:18 #link https://review.opendev.org/c/openstack/neutron/+/798009 14:24:21 But after looking at this patch, seems the author did not try to change the code like this. 14:24:57 It's only for the notify-callback mechanism in server side. 14:26:35 Let's continue the review of the patch. 14:26:40 OK, no more bugs 14:26:53 #topic L3_RFEs 14:27:04 #link https://bugs.launchpad.net/neutron/+bug/1931953 14:27:42 This is the spec for this RFE https://review.opendev.org/c/openstack/neutron-specs/+/796746 14:28:11 I've done some review works. it seems that this is still in a very begining state. 14:29:48 A complete description of the refactoring of this may require thousands of lines of spec, DVR is really a complex design. 14:31:00 Since this RFE was approved, so hope we can achive this goal someday. 14:32:02 #link https://bugs.launchpad.net/neutron/+bug/1933222 14:32:27 We have done the work of distributed DHCP: https://review.opendev.org/q/topic:%22bp%252Fdistributed-dhcp-for-ml2-ovs%22+(status:open%20OR%20status:merged) 14:33:01 So it's time for us to say good by to metadata-agent in the future. 14:33:35 Less neutron agents, more large scale for the cloud. 14:33:55 So in our local cloud deployment, we offline the DHCP-agent and metadata-agent. 14:34:57 After we have alternatives to DHCP-agent and metadata-agent. 14:36:31 Alright, no more things from me now. 14:38:17 #topic On demand agenda 14:41:05 Hi 14:41:46 Can someone point me to a resource where it describes what happens in the neutron in the background when a network is created and also when a vm is booted ? 14:43:40 OK, seems to be a one person meeting. 14:43:50 Let's end here. 14:43:53 Bye 14:43:55 #endmeeting