14:03:16 <liuyulong> #startmeeting neutron_l3
14:03:17 <openstack> Meeting started Wed Sep  9 14:03:16 2020 UTC and is due to finish in 60 minutes.  The chair is liuyulong. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:03:23 <openstack> The meeting name has been set to 'neutron_l3'
14:03:49 <liuyulong> Hi there
14:05:07 <liuyulong> #topic Announcements
14:05:33 <liuyulong> Seems another one person meeting, : )
14:06:43 <liuyulong> #link http://eavesdrop.openstack.org/meetings/networking/2020/networking.2020-09-08-14.00.log.html#l-12
14:06:57 <liuyulong> Let's recall the team announcments.
14:07:43 <liuyulong> Seems there are some L3 related works will not be able to added to Victoria cycle.
14:08:47 <liuyulong> The time is running really fast, I just remember that we had the V PTG a few weeks ago.
14:09:03 <liuyulong> And now it will be done then
14:10:25 <liuyulong> So any BPs/SPECs/RFEs which were not merged should be move to the W folder after the feature freeze.
14:10:57 <liuyulong> #link https://etherpad.opendev.org/p/neutron-wallaby-ptg
14:11:39 <liuyulong> And the PTL is gathering the topic for W vitual PTG, if you have, go ahead and write it down. : )
14:12:35 <liuyulong> OK, actually no announcements from me, just some repeats...
14:13:34 <liuyulong> OK, let's move on
14:14:36 <liuyulong> #topic Bugs
14:15:43 <liuyulong> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-September/017054.html
14:16:03 <liuyulong> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-September/016924.html
14:16:20 <liuyulong> The last two weeks bug list from our deputy.
14:17:04 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1892846
14:17:05 <openstack> Launchpad bug 1892846 in neutron "AttributeError in l3-agent changing HA router to non-HA" [Medium,In progress] - Assigned to Slawek Kaplonski (slaweq)
14:18:31 <liuyulong> Seems a race condition in agent side between changing the router HA-attribute (added) and removing the old one.
14:19:01 <liuyulong> #link https://review.opendev.org/#/c/747922/
14:20:39 <liuyulong> A fix have been made by slaweq, so we can move to there.
14:22:15 <liuyulong> IIRC the router delete action will bypass all the other router processing action by default.
14:22:58 <liuyulong> For one router, it can not be run in deleting and adding/updating in parallel.
14:24:26 <slaweq> hi
14:24:56 <slaweq> this patch is kind of WIP because I need to address ralonsoh's comments there
14:25:28 <liuyulong> So maybe the race condition I said upper is not right. We could revisit the code path to make sure that.
14:25:45 <liuyulong> slaweq, ack
14:26:53 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1893627
14:26:55 <openstack> Launchpad bug 1893627 in neutron "For neutron-l3-agent, after the execution of the linux command fails, it is not displayed which command failed to execute" [Wishlist,Fix released] - Assigned to boluan ace (boluanace)
14:27:05 <liuyulong> The fix https://review.opendev.org/#/c/749076/ is merged.
14:27:37 <liuyulong> It will be useful for developers when their external command get failed.
14:28:23 <liuyulong> OK, no more L3 bugs from the deputy's list. Let's revisit some old ones.
14:28:43 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1880532
14:28:44 <openstack> Launchpad bug 1880532 in neutron "[RFE]L3 Router should support ECMP" [Wishlist,New] - Assigned to XiaoYu Zhu (honglan0914)
14:28:55 <liuyulong> I noticed that ZhuXiaoYu is here.
14:29:00 <liuyulong> ZhuXiaoYu, hi
14:29:30 <ZhuXiaoYu> hi
14:30:51 <liuyulong> The rendered doc looks good to me now.
14:30:52 <ZhuXiaoYu> I just uploaded the functional test case today.
14:30:54 <liuyulong> #link https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_7db/729532/30/check/openstack-tox-docs/7dbc2e2/docs/specs/victoria/l3-router-support-ecmp.html
14:31:36 <ZhuXiaoYu> glad to hear that
14:32:44 <liuyulong> ZhuXiaoYu, I have one suggestion, it's better to add some OSC command examples to illustrate the user workflow.
14:34:58 <ZhuXiaoYu> Sorry, but what is OSC command?
14:35:06 <liuyulong> Sorry I lost the connetion
14:35:17 <liuyulong> "openstack client"
14:36:17 <liuyulong> The openstack client command examples can be the enrichment of the spec which will also tell the user that you can directly enable ECMP routing it by the exsitings.
14:36:47 <ZhuXiaoYu> Does the extra route has OSC?
14:37:50 <liuyulong> Yes
14:38:14 <liuyulong> Please try this "openstack router set --route destination=<subnet>,gateway=<ip-address> <router>"
14:39:06 <ZhuXiaoYu> Thanks for your advice, I will add that part later
14:39:47 <liuyulong> ZhuXiaoYu, cool
14:40:12 <liuyulong> Next one is from me
14:40:13 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1817881
14:40:14 <openstack> Launchpad bug 1817881 in neutron " [RFE] L3 IPs monitor/metering via current QoS functionality (tc filters)" [Wishlist,In progress] - Assigned to LIU Yulong (dragon889)
14:40:49 <liuyulong> This has the spec https://review.opendev.org/#/c/658511/
14:40:56 <liuyulong> and the POC https://review.opendev.org/#/c/675654/
14:41:31 <liuyulong> But seems no feedbacks after the RFE approvement.
14:42:15 <liuyulong> It should be move to W probably, : )
14:42:37 <liuyulong> Last one
14:42:38 <liuyulong> #link https://launchpad.net/bugs/1877301
14:42:40 <openstack> Launchpad bug 1877301 in neutron "[RFE] L3 Router support ndp proxy" [Wishlist,In progress] - Assigned to yangjianfeng (yangjianfeng)
14:42:42 <liuyulong> #link https://review.opendev.org/#/c/728628/
14:42:54 <liuyulong> The SPEC ^^
14:43:12 <liuyulong> The related API and extension is fine to me.
14:43:36 <liuyulong> But I have pointed some technical problems on data plane.
14:44:45 <liuyulong> It's about the GARP actions in IPv6 angle of view, aka the RA should be send out when the VM is migrating or HA state changing.
14:45:51 <liuyulong> It's more and more like a new floating IPv6 resource.
14:47:10 <liuyulong> Everthing is aligning with floating IP gradually.
14:47:43 <liuyulong> OK, no more bugs from me now.
14:47:51 <slaweq> I have one old one to mention
14:48:00 <liuyulong> Sure, go ahead
14:48:03 <slaweq> actuallyu patch for review: https://review.opendev.org/#/c/601336/
14:48:27 <slaweq> I just wanted to say that I made it configurable now and it seems it is ready for review and working
14:48:37 <liuyulong> Yes, I noticed that.
14:48:50 <liuyulong> Thank you for adding that.
14:49:18 <slaweq> thx a lot
14:49:28 <slaweq> I spent a lot of time on it recently
14:50:14 <liuyulong> Yes, it is complicated.
14:51:15 <liuyulong> Upstream CI will booting the real Ubuntu/CentOS image to run the keepalived?
14:51:24 <liuyulong> s/boot
14:51:58 <liuyulong> If not, the test may not be easy to add and verify.
14:52:02 <slaweq> no, u/s ci for now don't have any test for this new behaviour
14:52:18 <slaweq> but test should be easy to do in multinode dvr env
14:52:33 <slaweq> I was locally just adding new IP address in the vm and send garp using arping tool
14:54:32 <liuyulong> OK, I will re-test it if I have enough time. : )
14:54:39 <slaweq> thx a lot liuyulong
14:59:07 <liuyulong> OK, let's end here.
14:59:11 <liuyulong> Thank you guys.
14:59:15 <liuyulong> bye
14:59:19 <liuyulong> #endmeeting