16:15 <imaximets> o/

16:15 <mkalcok> \o

16:16 <imaximets> Mark said he'll not be able to join today.

16:16 <imaximets> So, I'll try and host.

16:17 <mkalcok> +1

16:17 <fnordahl> o/

16:17 <imaximets> Hi. It's time for a weekly OVN community meeting.

16:17 <imaximets> As I said, Mark won't be here today, so I'll host.

16:17 <imaximets> And I'll start with my report.

16:18 <imaximets> I don't have much. We officially released OVS 3.4 last week.

16:18 <imaximets> And I prepared a small patch to move OVN to v3.4.0.

16:18 <imaximets> *submodule

16:19 <imaximets> I think, Mark was planning to make a series of stable releases.

16:19 <imaximets> He posted patches, but didn't make actual releases yet, AFAIK.

16:19 <imaximets> That's all I have.

16:20 <imaximets> Who wants to go next?

16:20 <mkalcok> I can go next

16:20 <imaximets> mkalcok, sure.

16:20 <mkalcok> I had two things this week

16:22 <mkalcok> While putting the new "routing-protocol-redirect" option trhough its paces, We found out that there's pottentially an issue with BGP unnumbered (at least with FRR) as it may rely on some additional signalization via arp/ND that we are blocking to avoid duplicate RAs/ARP replies

16:22 <mkalcok> investigation is underway

16:23 <imaximets> ack

16:23 <mkalcok> I also noticed patch that amusil posted regarding the FIPs and SNAT with wide logical address pools. I'll put some time lookig at it as well

16:23 <mkalcok> that's it from me

16:24 <imaximets> mkalcok, thanks!

16:24 <zhouhan> Thanks mkalcok. I have been following up with this feature. Other than the "redirect" patch commited, is there any overall design documentation for this?

16:25 <fnordahl> zhouhan: there is the RFC thread on ovs-dev

16:25 <zhouhan> fnordahl: thanks, I will revisit that one

16:25 <zhouhan> I have no update other than some follow-up reviews. Currently on _lore_'s bfd lookup optimization series, and will send feedback/ack today.

16:26 <fnordahl> zhouhan: https://mail.openvswitch.org/pipermail/ovs-dev/2024-August/416296.html it has some instances in other months as well

16:26 <zhouhan> thanks fnordahl

16:26 <imaximets> thanks, zhouhan !

16:26 <imaximets> Who else want's to share?

16:26 <fnordahl> I have a quick update

16:26 <imaximets> fnordahl, go for it!

16:27 <fnordahl> mostly downstream stuff, but we put ovs 3.4.0 release and ovn snapshot into the development release of Debian/Ubuntu, during paces in CI we hit some flaky OVS tests, raised a patch here: https://mail.openvswitch.org/pipermail/ovs-dev/2024-August/416296.html

16:27 <fnordahl> Other than that I've been helping out with testing as per the discussion mkalcok raised above

16:27 <fnordahl> That's about it for me

16:28 <imaximets> fnordahl, thanks!

16:28 <imaximets> fnordahl, kernel in Ubuntu 24.04 seems to miss some conntrack fixes as well. :)

16:29 <imaximets> Causing FTP orig tuple tests to fail.

16:29 <imaximets> fyi

16:29 <fnordahl> imaximets: one of my colleagues approached me with an issue like this, do you have a pointer to them?

16:30 <imaximets> fnordahl, the following commit seem to be missing in the kernel: a23ac973f67f ("openvswitch: get related ct labels from its master if it is not confirmed")

16:30 <fnordahl> imaximets: thanks a bunch, I'll have a look

16:31 <imaximets> One other big thing for OVN 24.09 is that there is a revert for ECMP nexthop ct flushing feature on the way.

16:31 <imaximets> _lore_ posted today.

16:32 <imaximets> It flushes too much and it seems that there is no easy way to fix it in 24.09.

16:32 <_lore_> imaximets: yes, the solution it needs a rework, so better to delay it

16:33 <imaximets> _lore_, do you have some updates to share?

16:33 <_lore_> imaximets: do I need to post v2 of the revert to update the db version?

16:33 <_lore_> not much, I posted the improvement we discussed with zhouhan

16:33 <imaximets> _lore_, should be easier this way, I guess.

16:34 <_lore_> and then I was looking at this bug

16:34 <_lore_> ok, I will post v2 soon

16:34 <imaximets> _lore_, thanks! Anything else?

16:34 <_lore_> nope, thx

16:34 <imaximets> ack

16:35 <imaximets> OK. Anyone else here wants to share?

16:36 <imaximets> I'll take silence as a 'no'. :)

16:36 <imaximets> Thanks, everyone!

16:36 <imaximets> Bye.

16:36 <fnordahl> \o thanks and cheers!

16:36 <mkalcok> thanks o/

16:38 <zhouhan> thx & bye