16:15 <mmichelson> Hi everyone. Welcome to the weekly OVN developers' meeting.

16:15 <amusil> Hi

16:15 <mkalcok> \o hello

16:15 <mmichelson> I can give the first update.

16:16 <mmichelson> The only real upstream activity I have to report is that I put up a patch that adds documentation for the "--portrange" option for `ovn-nbctl lr-nat-add` to the ovn-nbctl manpage.

16:16 <mmichelson> I've also done some reviews.

16:16 <mmichelson> I think that's about all I have to report. So who would like to go next?

16:17 <amusil> I can

16:17 <mmichelson> go for it amusil

16:18 <amusil> We've got a report from Openshiftthat tunneled traffic has wrong encap src ip with multiple IPv6 addresses on the interface

16:19 <amusil> To prevent that we would like to always enforce the local_ip option

16:20 <amusil> The removal of "ovn-set-local-ip" seems straight forward however there are some issues with dummy datapath I'm looking into that

16:20 <amusil> Besides that there is v2 of AS I-P patch, that's it

16:20 <mmichelson> Thanks amusil . Who would like to go next?

16:21 <mkalcok> I can go next

16:22 <mkalcok> Working on v4 for Direct SNAT access patch. Thanks for reviews amusil, dceara. After reusing parse/encode functions from ct_commit_nat, as requested, I'm seeing first packet not getting delivered. I suspect that this stems from differences between explicit "next;" action (that I used in v3) and implicit "next" provided by ct(table=X) that I'm using now.

16:22 <mkalcok> So I'm trying to figure out how to fix this. That's it from me.

16:23 <mmichelson> Thanks mkalcok . Anyone else?

16:23 <fnordahl> super short update

16:24 <fnordahl> I've been looking into a aparently unfixable regression in branch-22.x which is isolated to ICMP packets to lb in specific config, thanks for responding on the question about it amusil. Looks like we've settled on awaiting upgrade to 24.03 soon

16:24 <fnordahl> that's it for me

16:25 <amusil> It would be possible to backport it's just matter of considering if that's really worth it

16:25 <fnordahl> amusil: yup, I tend to agree

16:25 <mmichelson> Thanks fnordahl . Does anyone else have an update?

16:25 <fnordahl> amusil: our rep first thought it was also impacting traffic, but once we realized it was icmp only I think we've found a pragmatic middle ground

16:26 <amusil> Yeah that was one of the reasons why we didn't "bother" with backport

16:26 <amusil> It shouldn't have any huge effect on real traffic

16:28 <mmichelson> OK, I guess nobody else wants to give an update?

16:29 <zhouhan> I was mostly on internal stuff, so no update for last week. I will get some time with amusil's I-P patch this week.

16:29 <mmichelson> Sounds good zhouhan

16:29 <amusil> Thanks

16:31 <mmichelson> OK, I think that's all for this meeting then. Thanks eveyrone and have a good day!

16:31 <fnordahl> \o cheers!

16:31 <mmichelson> Bye!