Hello all, been suffering with a really weird issue and can't seem to uncover what is going on. Each time I reboot my server, only SOME of my docker containers will start, it seems those using br0 all fail with

Mar 28 12:44:13 Server rc.docker: nextcloud: Error response from daemon: failed to create the ipvlan port: device or resource busy 
Mar 28 12:44:13 Server rc.docker: Error: failed to start containers: nextcloud 
Mar 28 12:45:04 Server kernel: br0: Device is already in use.

The only fix that works for me is, to disable the docker service, disable "Host access to custom networks", Apply, then enable "Host access to custom networks" and start the docker service, and then the containers start fine. I am using IPVLAN. Not sure where to go from here as there are no other log entries for it and googling produces no results. Any help? It's very frustrating having to do this on EVERY reboot of the server, and it seems like such a simple problem.

ogreserv-diagnostics-20240328-2036.zip

Any idea? This is happening every single reboot. Just happened today, only workaround that works is Docker service > Disable, host access > disable, host access> enable, docker > enable.

 

This is severely frustrating to have to do every single reboot. Please help lol

Try uninstall Tailscale and check again.

On 4/12/2024 at 12:19 AM, Vr2Io said:

Try uninstall Tailscale and check again.

Sorry for the late reply. Uninstalling Tailscale did not fix the issue. To add to the information, I just updated to 6.12.10 and the issue persists, same workaround is necessary each reboot still.

Okay so, after working on a completely separate project involving a docker networking setting, I ran across something that helped me to resolve this.

TURNS OUT, back in version 6.9.x, there was a docker issue I was having relating to networking, and I had placed a USER SCRIPT to manually add the docker network at boot time. Well, this configuration used a completely different IP range than what the server is using now. I completely forgot about this script existing or to even check for it, so my ENTIRE issue was caused by this rouge script I had placed. After disabling/deleting the script and rebooting, all docker containers boot up perfectly fine!

 

Turns out, the reason toggling host access and restarting docker worked, it was just deleting the docker network, and recreating it as it was supposed to be. I feel silly but, HEY, IT'S FIXED!

Edited June 24, 20242 yr by TheOgre