16:15 <mmichelson> Hi everyone. It's time to start the weekly OVN developers' meeting.

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

16:16 <mmichelson> I posted a new version of the datapath and port binding sync refactor. Mirror port support was added to OVN main since the last version I posted, so I adjusted the code to work with mirror ports.

16:16 <mmichelson> I also did some reviews.

16:17 <mmichelson> I'm now working on something similar (but much less drastic) for logical flows. The nice thing about logical flows is that they are a final product. Nothing uses logical flows as an input, so the changes are going to be more incremental and hopefully much easier to review.

16:18 <mmichelson> That's all from me. Who would like to go next?

16:18 <fnordahl> o/

16:19 <mmichelson> Hi fnordahl . Do you have an update to give?

16:19 <fnordahl> Sure

16:19 <fnordahl> Sprinting this week, but posted a fix for a test that fails only on s390x due to assumptions on the first bucket being used for load balancing to multiple backends

16:20 <fnordahl> Working on downstream integration on BGP stuff, and found a potential bug in the route lerning code related to multiple next hops, so will continue to investigate that

16:20 <fnordahl> That's about it for me

16:20 <mmichelson> fnordahl, thanks. Is the test failure consistent or does it only happen sometimes?

16:20 <fnordahl> mmichelson: consistent

16:21 <mmichelson> fnordahl, ack, I'll have a look at the review soon.

16:21 <fnordahl> ty!

16:21 <mmichelson> s/review/patch/

16:21 <mmichelson> Does anyone else wish to give an update? I'm not sure who all is actually here.

16:22 <zhouhan> Hi

16:22 <mmichelson> dceara said he'd be late. amusil is off today. imaximets is at a conference.

16:22 <mmichelson> zhouhan, hi!

16:22 <zhouhan> I am still working on the commit-all fix for HW offload, distracted by other issues

16:23 <zhouhan> I have a version that fixed the HW offload, verified with HW

16:23 <mmichelson> "distracted by other issues" is my entire working life I feel like.

16:23 <zhouhan> but it breaks some test cases - it is really hard to verify this, because currently the commit-all is an option and not the default

16:25 <zhouhan> So I had to manually change the code to set it as default and run test cases, and many would fail just because the lflows and ct entries are not as expected. But I checked some specific test scenarios and found that there are actually functional failures. So I am still debugging

16:26 <zhouhan> "distracted by other issues"  is my entire working life I feel like. -- mmichelson: yes, one of the issue is OVS handling for ECN bits which caused HW offload failure ...

16:26 <zhouhan> anyway, that's it for me

16:26 <mmichelson> thanks zhouhan

16:27 <mmichelson> dceara, I think you're the only person here, so feel free to share your update.

16:27 <dceara> mmichelson, ack

16:27 <mmichelson> s/only/last/

16:28 <dceara> I don't have much this week (it's a period with random holidays here). Maybe the only interesting thing is that I've been trying to outline a plan for enhancing the dynamic routing support for BGP+EVPN. I've opened some tasks for that in our internal (Red Hat) tracking system. I'll speak about it a bit during the next technical meeting.

16:29 <dceara> Which reminds me, we have our next technical meeting scheduled for Monday next week for those who can join.

16:29 <dceara> I'll send a reminder on-list tomorrow too.

16:29 <mmichelson> I plan to be there.

16:29 <dceara> That's it on my side.

16:29 <dceara> Thanks!

16:29 <fnordahl> Great to see it being builded upon ++

16:29 <mmichelson> Thanks dceara!

16:29 <fnordahl> Our entire team is sprinting next week and we plan to attend the A/V meeting from here

16:29 <mmichelson> Just to double-check, is there anyone else here that wants to give an update?

16:31 <mmichelson> OK I think we're done.

16:31 <mmichelson> Have a great day everyone!

16:32 <mmichelson> Bye!

16:32 <dceara> Thanks, bye!

16:32 <fnordahl> \o thx, cheers!