14:00:41 #startmeeting neutron_l3 14:00:41 Meeting started Wed May 20 14:00:41 2020 UTC and is due to finish in 60 minutes. The chair is liuyulong. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:42 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:45 The meeting name has been set to 'neutron_l3' 14:01:11 hi 14:01:24 hi 14:01:41 Hey! 14:02:09 Hi 14:02:12 OK, let's start 14:02:13 #topic Announcements 14:03:00 hi 14:03:18 #link https://etherpad.opendev.org/p/neutron-victoria-ptg 14:03:48 Only a few days, we will begin the V virtual ptg online. 14:05:33 Maybe I should add something that was mentioned in Shanghai for a further discussion. 14:05:48 #link https://doodle.com/poll/xcd82ea6svs5dben 14:06:31 From the time slot here, I should have only 2 hours to meet you guys all at the same time online. 14:06:43 Each day 14:07:32 Anyway, again topics are welcomed to be added to the list. 14:08:17 More announcements please have a look at: 14:08:18 #link http://eavesdrop.openstack.org/meetings/networking/2020/networking.2020-05-19-14.00.log.html#l-12 14:08:25 OK, no more from me then. 14:10:10 OK, let's move on 14:10:18 #topic Bugs 14:10:35 #link http://lists.openstack.org/pipermail/openstack-discuss/2020-May/014913.html 14:10:49 The bug list from our deputy last week. 14:11:02 First one 14:11:05 #link https://bugs.launchpad.net/neutron/+bug/1877977 14:11:05 Launchpad bug 1877977 in neutron "[DVR] Recovery from openvswitch restart fails when veth are used for bridges interconnection" [Medium,Confirmed] 14:12:23 I want to mention one more action 14:13:19 change the config option and run "kill -SIGHUP ", this should help to recovery the dataplane? 14:14:02 I think the idea was to deprecate veth pairs for bridge connectivity 14:14:09 and use only patch ports 14:15:44 And typically, when the openvswitch (vswitchd) is restarted, the canary flows will be lose. 14:16:07 The ovs-agent monitor this and will consider it should run a full sync. 14:16:27 Everything should be added back again. 14:17:14 If not, this should be a bug absolutely. 14:17:18 it should, by the ovs agent 14:18:33 More information about this is we use the systemd to do the next step from the bug "restart the neutron-ovs-agent." 14:19:16 If the openvswitch is restarted, the ovs-agent will run restart action automatically. 14:20:19 OK, next one 14:20:41 #link https://bugs.launchpad.net/neutron/+bug/1878299 14:20:41 Launchpad bug 1878299 in neutron "[floatingip port_forwarding] changing external port to used value hangs with retriable exception" [Medium,Confirmed] 14:21:52 I did not reproduce the bug locally, but haleyb has confirmed the bug. 14:23:15 Looks like some DB related issue when the 'port' check is running. 14:24:28 I need to check with haleyb what is really happening and how I can reproduce it 14:24:40 in order to try code a fix 14:27:03 ralonsoh, thank you. I just run a quick test locally, not see the such issue. 14:27:05 ralonsoh: hi, sorry, in another meeting 14:28:42 OK, let's try to dig this locally in details. 14:28:47 Next 14:29:59 #link https://bugs.launchpad.net/neutron/+bug/1877301 14:29:59 Launchpad bug 1877301 in neutron "[RFE] L3 Router support ndp proxy" [Wishlist,Confirmed] 14:30:22 We have discussed this in last L3 meeting, and it was approved during last drivers meeting. 14:30:59 yangjianfeng has submitted the spec 14:31:01 #link https://review.opendev.org/#/c/728628/ 14:31:40 So we can continue the discussion in this spec. 14:32:17 Next one: 14:32:19 #link https://bugs.launchpad.net/neutron/+bug/1879215 14:32:19 Launchpad bug 1879215 in neutron "[L3] Unexcepted HA router scheduled instance shows up after manully scheduling" [Undecided,New] 14:33:00 This was reported by me, it is easy to reproduce. 14:34:10 These action is useful when the operators want to move the network node offline. 14:34:28 Looks like we're relying on some old data during --enable operation. 14:34:56 The existing router resources (scheduled routers) should be moved from one to another. 14:36:21 The priority of this bug isn't to big, right? I could try to reproduce it and triage. 14:38:44 maciejjozefczyk, yes, it's not big. But it breaks the config limitation. 14:38:47 my colleague found the problem is related to this function. https://github.com/openstack/neutron/blob/master/neutron/db/l3_agentschedulers_db.py#L507 14:39:29 maciejjozefczyk, I will be cool if you can confirm if it is also existing in master branch. 14:39:51 d/I 14:40:46 I will submit the fix after we check and test it really works locally. 14:40:50 liuyulong, yes I'll try to check if this exists in master. 14:40:57 ok, liuyulong 14:41:40 OK, no more bugs from me now. 14:41:44 Any updates? 14:42:14 * liuyulong scanning the LP bug list... 14:43:37 OK, I see something related to L3. 14:43:40 #link https://bugs.launchpad.net/neutron/+bug/1878905 14:43:40 Launchpad bug 1878905 in neutron "when I use dvr mode for neutron,it's not functional" [Undecided,Incomplete] 14:44:24 Yeah, the bug is not clear to me. Looks like the guy deployed the env without DVR and then switched it to DVR? Hard to tell... 14:44:29 maciejjozefczyk has marked it as incomplete. Yes, I will do same thing. 14:45:11 maciejjozefczyk, exactly, I have the same question. 14:46:05 yeah, what about the bridge configuration on computes, perhaps its missing, I dunno. 14:46:22 And the config option should be pasted to the bug for the team to analyse. 14:46:36 ++ 14:46:38 But as per Kolla cores they have the test covering DVR scenario. 14:46:49 I'll ask for inputs from computes and neutron config 14:47:59 Yes, we need some clue to do the trouble shooting, not guess... : ) 14:48:07 ;) 14:48:47 #link https://bugs.launchpad.net/neutron/+bug/1877978 14:48:47 Launchpad bug 1877978 in neutron "SNAT Problem Floating IP" [Undecided,Incomplete] 14:49:18 This should be another "unfriendly bug report"... 14:49:23 :) 14:50:06 I have such encounter such phenomenon, you guys just guess... 14:51:13 #link https://docs.openstack.org/neutron/pike/contributor/policies/bugs.html#bug-report-template 14:51:46 I guess for the bug deputy, you guys could easily paste this link to the bug. 14:52:26 liuyulong++ ok, I'll paste 14:52:43 Sorry, it is not the latest. 14:52:51 * liuyulong is looking for it now... 14:53:25 #link https://docs.openstack.org/neutron/latest/contributor/policies/bugs.html#bug-report-template 14:54:34 OK, no more from me. (really this time) 14:56:35 Actually no OVN related bugs were reported in last 2 weeks, so let's move to the last topic (we are running out of time). 14:56:42 #topic On demand agenda 14:57:50 mlavalle, FYI, yangjianfeng has submitted the spec of the IPv6 ndp proxy: https://review.opendev.org/#/c/728628/ 14:58:23 liuyulong: ack. Thanks 14:58:34 in my pile 14:58:59 liuyulong: does he work with you? 14:59:21 mlavalle, actually yangjianfeng is my former colleague. He has moved to a new position in another company. 14:59:44 liuyulong: ack 15:00:10 For the ipv6 ndp_proxy, in china telecom we have implemented something similar to this. 15:00:49 But we introduce a new resource, IPv6Address (something like floating IPv6), to routers to enable the ndp_proxy 15:01:37 It is mainly for productization of our public cloud 15:01:43 OK, time is up... 15:01:54 Thank you guys. 15:01:57 #endmeeting