I am running Unraid 6.12.4. I appear to have a networking issue - network br0 not found. A couple of dockers that use Custom IP addresses won't start. The either give "Execution Error - No Such Container" or "Execution Error - Bad paramater".

Here is a snippet from my log file: Nov 27 19:31:29 Portrush rc.docker: sagetv-opendct-java11: Error response from daemon: network br0 not found

How do I fix this?

I found a post from 2018 - https://forums.unraid.net/topic/68468-solved-missing-br0-interface-in-gui/ where @bonienl says:

rm /var/lib/docker/network/files/local-kv.db
/etc/rc.d/rc.docker restart

Would that still work? In order to do that command do I have to stop the docker service, or can I do it with the docker service running?

portrush-diagnostics-20251127-1832.zip

I would recommend starting by updating Unraid to the latest stable version or at least to 6.12.15 if you want to remain on 6.12 for some reason.

Ok - is there anything to be aware of if I am going to upgrade to 7.2.2? I haven't upgraded recently as I haven't sent much time playing around with this system. I have no reason to want to be on 6.12 or any other release.

When I go into Update OS it shows updating to 7.0.1. Do I have to update to that version first, and then go to 7.2.2?

You should be able to go directly to 7.2, but if it doesn't show, update to 7.0 first, create a flash drive backup prior to updating. Also, recommend reading the release notes for the major ones, like 7.0.0, 7.1.0 and 7.2.0

I have upgraded to 7.2.2 with an interim momentary upgrade to 7.0.1. But I still seem to be having the same error - although it is giving a different name for the network device that isn't foundportrush-diagnostics-20251128-1122.zip. In the settings for the dockers that are failing the network type is currently shown as Custom: br0. I want these two dockers to get their own IP addresses.

Should I now let docker rebuild the network with the commands shown above in my OP? Is that deprecated (it was from 2018). Is there another way to give those dockers their own IPs these days?

This is from my log:

Nov 28 11:20:39 Portrush rc.docker: pihole: Error response from daemon: network 5d38b8bd42b3d46eb1e75824772666a658bfd1c7b8f1d89e03a5981c9dee4212 not found
Nov 28 11:20:39 Portrush rc.docker: Error: failed to start containers: pihole
Nov 28 11:20:39 Portrush rc.docker: sagetv-opendct-java11: Error response from daemon: network 5d38b8bd42b3d46eb1e75824772666a658bfd1c7b8f1d89e03a5981c9dee4212 not found
Nov 28 11:20:39 Portrush rc.docker: Error: failed to start containers: sagetv-opendct-java11

Attach a new diagnostics file

Docker networking is not really my forte, but you can try changing to ipvlan to see if the network is recreated correctly; that is the default now, though you could change back to macvlan if that works.

Another option would be going to the containers that are using a custom br0 assignment, changing to bridge, applying, changing back to custom br0, setting IP, and applying again.

Thanks, I have it working now. First I switched to IPVLAN. That didn't fix it. I kept it on that setting and switched the dockers to Bridge, and then to br0 without actually saving. When I then applied the change they started working.

Should I keep going on IPVLAN, assuming that it doesn't cause any other issues? Or switch back to MACVLAN? Anything to keep an eye out for with a different network type?

Ipvlan is recommended for all users, unless you miss something from Macvlan, I believe the main difference is that Macvlan will get a different MAC address for the containers, while they all get the same one with ipvlan, which is typically not a concern for most use cases.