13:58:47 <liuyulong> #startmeeting neutron_l3
13:58:48 <openstack> Meeting started Wed Feb 12 13:58:47 2020 UTC and is due to finish in 60 minutes.  The chair is liuyulong. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:58:49 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:58:51 <openstack> The meeting name has been set to 'neutron_l3'
13:58:57 <liuyulong> #topic Announcements
13:59:10 <slaweq> hi
13:59:30 <liuyulong> hi
13:59:43 <liuyulong> Ussuri-2 is this week.
14:00:15 <liuyulong> And we have this list:
14:00:17 <liuyulong> #link https://launchpad.net/neutron/+milestone/ussuri-2
14:02:01 <liuyulong> Two RFE related bugs are not finished. So maybe we should move it to next U-3
14:02:42 <liuyulong> OVN migrations are quite good I guess.
14:03:23 <liuyulong> OK, I just read the book, haha.
14:03:29 <liuyulong> No more from me now.
14:03:39 <liuyulong> Any updates?
14:04:12 <liuyulong> OK, let's move on.
14:04:15 <liuyulong> #topic Bugs
14:04:24 <ralonsoh> hi
14:04:35 <liuyulong> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-February/012477.html
14:04:38 <liuyulong> ralonsoh, hi
14:05:23 <liuyulong> First one
14:05:25 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1862315
14:05:26 <openstack> Launchpad bug 1862315 in neutron "Sometimes VMs can't get IP when spawned concurrently" [High,New] - Assigned to Oleg Bondarev (obondarev)
14:05:58 <liuyulong> It has a fix now.
14:06:00 <liuyulong> #link https://review.opendev.org/#/c/707077/
14:09:08 <liuyulong> sorry, a bit latency on networking...
14:09:31 <liuyulong> Yes, the fix is fine to me, we can go with that.
14:11:57 <liuyulong> We have that original fix https://review.opendev.org/#/c/626830/ in our queens branch for a long time, in general it works fine.
14:13:11 <liuyulong> I will check our local logstash to see if it has such issue.
14:14:51 <liuyulong> OK, no more bugs in the deputy list, let's give a quick scan of the the LP.
14:15:38 <liuyulong> ++,sorry, no more bugs related to L3 in the deputy list
14:17:03 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1862851
14:17:04 <openstack> Launchpad bug 1862851 in neutron "update_floatingip_statuses: StaleDataError: UPDATE statement on table 'standardattributes' expected to update 1 row(s); 0 were matched." [Undecided,New]
14:17:37 <haleyb> hi, sorry was in a downstream chat
14:17:57 <ralonsoh> this is not the first time we have this bug
14:18:00 <liuyulong> This is a really common exception in DB related actions.
14:18:03 <haleyb> we've seen those errors before ^^^, and that bug is against rocky
14:19:05 <ralonsoh> btw, a patch for a previous bug related to this
14:19:06 <ralonsoh> https://review.opendev.org/#/c/503517/
14:19:10 <ralonsoh> https://bugs.launchpad.net/neutron/+bug/1716321
14:19:12 <openstack> Launchpad bug 1716321 in neutron "StaleDataError: UPDATE statement on table 'standardattributes' expected to update 1 row(s); 0 were matched." [Undecided,Fix released]
14:20:48 <liuyulong> Yes, we have met this sometimes also, as far as we can see the exception does not have big effect to the neutron-api, the DB retry action could work.
14:21:11 <ralonsoh> exactly
14:22:31 <liuyulong> OK, next
14:22:32 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1861876
14:22:33 <openstack> Launchpad bug 1861876 in neutron "[Neutron API] Neutron Floating IP not always have 'port_details'" [Critical,New]
14:22:57 <liuyulong> A critical one, but I did not read the detail yet...
14:23:49 <slaweq> I think it's fixed now (at least not so critical anymore)
14:23:59 <liuyulong> I don't think floating IP has such "port_details" attribute. It is another L3 service_providers ?
14:24:18 <ralonsoh> that was because OVN L3 pluging didn't have the correct mixin
14:24:34 <ralonsoh> the nova bug is solved
14:24:38 <ralonsoh> and neutron one
14:24:44 <ralonsoh> https://review.opendev.org/#/c/705982/
14:26:02 <liuyulong> Alright, maybe we can get it merged.
14:27:41 <liuyulong> OK, that's all from me.
14:27:53 <liuyulong> Any other bugs?
14:28:19 <slaweq> I want to ask You to review https://review.opendev.org/#/c/702856/
14:28:34 <ralonsoh> sure
14:28:44 <slaweq> liuyulong: I was playing with putting interface down on standby nodes but it didn't work as I wanted
14:29:13 <slaweq> there were some errors in keepalived logs during failover
14:29:59 <liuyulong> : ), maybe I should submit our local fix.
14:30:31 <slaweq> liuyulong: sure
14:30:42 <slaweq> if You have something which really works without problems than we can review it
14:30:51 <liuyulong> I will submit it today.
14:31:20 <slaweq> liuyulong: thx
14:33:03 <liuyulong> slaweq, np
14:33:07 <liuyulong> OK, let's move on
14:33:14 <liuyulong> #topic OVN_L3
14:34:15 <ralonsoh> maciejjozefczyk, ping
14:34:17 <ralonsoh> lucasagomes,
14:34:45 <ralonsoh> ^^ any relevant update this week about OVN L3?
14:34:58 <lucasagomes> hi there, not from my part
14:35:06 <liuyulong> We have a bug related to OVN gateway I remember.
14:35:06 <ralonsoh> btw, we are still dealing with the CI
14:35:31 <liuyulong> #link https://review.opendev.org/#/c/705660/
14:35:34 <liuyulong> This one
14:35:37 <lucasagomes> yeah, specially on the func tests... I know maciejjozefczyk is looking into it. He's uploaded a new patch-set today
14:36:16 <maciejjozefczyk> liuyulong, Yes we have an OVN gateway related patch: https://review.opendev.org/#/c/705660/
14:37:04 <maciejjozefczyk> I needs to be on top of functional one, and those functional test failed on something else.
14:37:17 <maciejjozefczyk> I pushed today a fix to stablize part of functional tests related failures
14:38:25 <liuyulong> Yes, I noticed that, it is based on 3 parts of functional test fix.
14:39:03 <maciejjozefczyk> liuyulong, yes, actually the functional test that I needed to change is in the third part.
14:40:02 <liuyulong> So I wonder if the stable branch could have the fix reviewing in parallel ?
14:40:35 <liuyulong> Stable branch does not need to fix the functional test, right?
14:41:31 <maciejjozefczyk> liuyulong, potentially we could do a review in a stable networking-ovn branch and then apply it on master, but that breaks workflow a bit ;/
14:42:06 <maciejjozefczyk> liuyulong, or we can review this patch now, have agreement on its shape, then I can cherry-pick it, what do you think?
14:42:24 <maciejjozefczyk> liuyulong, actually tests that are failing are unrelated.
14:44:53 <liuyulong> Both make sense for me, but IMO if it get merged first in networking-ovn, it will be more easily to backport to T/Q/R stable branches.
14:45:55 <liuyulong> For important patches, my idea is to quickly merge and benefits the end users.
14:46:58 <maciejjozefczyk> liuyulong, Ok, I'm gonna propose it in stalble/train first, we gonna have a discussion there. Then, if functional tests will be in place, we gonna merge it in Neutron master, agreed?
14:47:22 <slaweq> I think that we should propose in master first and then do "cherry-pick" to networking-ovn
14:47:41 <slaweq> I don't think we should change our normal workflow as maciejjozefczyk already said
14:49:21 <maciejjozefczyk> slaweq, liuyulong I'm ok with both, this change is important, but not a blocker.
14:49:26 <liuyulong> Sure, both works
14:50:02 <maciejjozefczyk> Would be great to have somebody eyes on it first :)
14:51:42 <liuyulong> #link https://bugs.launchpad.net/neutron/+bug/1862648
14:51:44 <openstack> Launchpad bug 1862648 in neutron "[OVN] Reduce the number of tables watched by MetadataProxyHandler " [High,Fix released] - Assigned to Lucas Alvares Gomes (lucasagomes)
14:52:44 <lucasagomes> yeah that's an optimization to the ovn metadata anget
14:52:46 <lucasagomes> agent*
14:52:46 <liuyulong> After reading this bug, I have an idea, if we can ovn-metedata-agent someday by using some distributed proxies locally in the compute node.
14:53:25 <liuyulong> sorry, my idea is to replace, "if we can replace"
14:54:55 <liuyulong> let the ovn-controller do something instead of a python agent.
14:56:35 <lucasagomes> it's something we could bring up with the OVN team itself, I've jsut heard of the idea so I don't know much about it atm
14:57:58 <liuyulong> Np, we have a local topic which is trying to replace python neutron-metadata-agent with distributed proxies locally.
14:58:45 <liuyulong> OK, time is up.
14:58:53 <liuyulong> Thank you guys, bye
14:58:56 <liuyulong> #endmeeting