Docker veth设备异常引发NetworkManager报错，系统卡顿求助

Hey there, let’s work through this annoying Docker veth device issue step by step—those repeated NetworkManager errors and constant interface flips are definitely causing that frustrating system lag you’re seeing. Here’s how to tackle each part of the problem:

1. Keep NetworkManager Away from Docker’s veth Interfaces #

NetworkManager tries to manage every network interface by default, which clashes hard with Docker’s own virtual network setup. Let’s tell it to ignore veth devices entirely:

• Create a config file at /etc/NetworkManager/conf.d/99-docker.conf with this content:

  [keyfile]
  unmanaged-devices=interface-name:veth*
  

• Restart NetworkManager to apply the change:

  sudo systemctl restart NetworkManager
  

This should stop the udev "device not found" and ifupdown configuration errors right away, since NetworkManager won’t waste time trying to configure interfaces it shouldn’t touch.

2. Validate Docker Network Configurations #

The "carrier detection not supported" warning for the veth driver is actually normal—veth interfaces don’t use physical carrier signals. But the frequent delete/recreate cycles might point to a wonky Docker network setup:

• List all your Docker networks to spot any custom ones that could be misbehaving:

  docker network ls
  

• Inspect suspicious networks to check for driver or subnet misconfigurations:

  docker network inspect <network-name>
  

If you’re using a custom network, try recreating it with Docker’s default bridge driver to rule out driver-specific bugs:

docker network rm <problem-network>
docker network create <new-network-name>

3. Check for Udev Rule Conflicts #

Sometimes custom udev rules can trigger unnecessary events for veth devices, causing them to flip unexpectedly:

• List network-related udev rules to spot any that target veth interfaces:

  ls /etc/udev/rules.d/*net* /lib/udev/rules.d/*net*
  

• If you find any rules that reference veth*, comment them out temporarily, then restart udev:

  sudo systemctl restart systemd-udevd
  

4. Update Docker & System Packages #

Outdated Docker versions or system libraries often cause compatibility issues with virtual network interfaces. Let’s get everything up to date:

• Update Docker to the latest stable version (adjust commands for your distro; example for Ubuntu):

  sudo apt update && sudo apt upgrade docker-ce docker-ce-cli containerd.io
  

• Update all system packages to patch any underlying bugs:

  sudo apt update && sudo apt full-upgrade
  

Reboot your system after updating to ensure all changes take effect.

5. Monitor for Resolved Errors #

After applying the fixes, keep an eye on the logs to confirm the issues stop:

• Watch NetworkManager logs in real-time:

  journalctl -u NetworkManager -f
  

• If problems persist, check Docker’s daemon logs for additional clues:

  journalctl -u docker -f
  

内容的提问来源于stack exchange，提问作者Mike Crowe