Any debugging tips for a Weird lxdbr0 issue with UB20 Containers with Centos7 host

I’ve started to run containers at home for key services (email and a nginx based services)…

My LXD host is a VMware Fusion node which is running Centos7… I’m running LXD4.9 and the containers are running Ubuntu 20.04.

I’m finding periodically both containers lose networking I’ve not done much debugging but if I restart the containers (lxc restart) then they regain their networking.

Any tips on debugging?

Can you confirm whether when they lose networking, do the containers lose their IP addresses, or do the network interfaces disappear or go down?

i.e. what is the specific nature of the problem? Also do the NICs use DHCP? And how long does it take for them to lose networking?

Finally, what form of networking are you using?

Next time it happens I’ll will gather some intel… Thanks

Quick update… I’ve not had any problems since last time… Which is 22days ago… So great…

Now the not so great think is I disabled IPV6 on the lxdbr0 and I ‘suspect’ thats what fixed the issue. Nothing else has changed… However this is a weak root cause as I did so on a whim.

In an ideal world I’d reenable it and offer todo debugging but I’m busy offloading the functionality to a RPI4 so we’ll see if it reoccurs on there…