[Solved] Bridge network hangs all network activity

[at88mph]

I'm running Debian 12.5 on a Dell Optiplex 990 with UEFI and a fast NVMe drive with 16Gb RAM. I've installed the latest Docker version 25. I have the exact same installation on a basic SATA spinning disk on the same machine as well.

By default, Docker uses the Bridge network (docker0 device) when running containers. I'm trying to build a Jekyll site using Docker, which fetches many Gem files from rubygems.org. My "docker run" command runs successfully for a short period of time, but then can no longer make network connections out. Also, my entire desktop's network capabilities are halted (i.e. browser, apt, etc.). When the container dies eventually, all network is restored on the machine. This is on my NVMe drive. If I specify the "--network host" option, it succeeds.

With the SATA drive installation, this all succeeds as expected. The only difference between the two installations is that the 8021q module (VLAN) is loaded by default in my NVMe install that might be getting in the way but I can't disable it.

I tested this with the latest Ubuntu and it works fine on that same NVMe disk.

Any thoughts on this? I use Docker very regularly.
Thank you, community!

[at88mph]

This ended up being an issue with connman, if anyone else runs into this. It was trying to configure bridge networks that Docker creates on the fly, so you must Blacklist them in the /etc/connman/main.conf file by uncommenting the `NetworkInterfaceBlacklist` entry. If that does not work, as it did not for me, I disabled connman, and replaced it with NetworkManager which did solve it.

https://forums.docker.com/t/bridge-netw ... ine/140212

[Aki]

Thank your for sharing the solution to the reported issue.

Happy Debian.