WSL-Docker-CudaGL-ROS2-Gazebo

🛠 Troubleshooting

This document provides solutions to common issues encountered when setting up ROS 2 Iron + Gazebo Garden + CUDA in Docker (WSL2).

1️⃣ nvidia-smi Not Detecting GPU Inside Docker

Issue

Running nvidia-smi inside the container returns:

Failed to initialize NVML: Driver/library version mismatch

or

NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver

Solution

  1. Ensure NVIDIA drivers are installed on the host (Windows): 
    wsl --shutdown

    Then, update drivers from:

  2. Ensure Docker has access to GPU: 
    docker run --rm --gpus all nvidia/cuda:12.2-base nvidia-smi
    • If this fails, reinstall nvidia-container-toolkit inside WSL: 
      sudo apt update
sudo apt install -y nvidia-container-toolkit

2️⃣ OpenGL Renderer is Not NVIDIA

Issue

Running glxinfo -B inside the container shows:

OpenGL renderer string: llvmpipe (LLVM 13.0.0, 256 bits)

instead of NVIDIA.

Solution

  1. Ensure libgl1-mesa-glx and Vulkan drivers are installed inside WSL: 
    sudo apt install -y libgl1-mesa-glx libvulkan1 mesa-vulkan-drivers vulkan-tools
  2. Restart WSL: 
    wsl --shutdown
  3. Verify again: 
    glxinfo -B

    It should now show:

    OpenGL renderer string: NVIDIA

3️⃣ Gazebo GUI Does Not Appear

Issue

Running gz sim -v 4 does not open the GUI.

Solution

  1. Ensure WSLg is enabled in WSL: 
    echo $DISPLAY

    If empty, restart WSL:

    wsl --shutdown
wsl
  2. Ensure docker-compose.yml has the correct X11 and WSLg environment: ```yaml environment:
    • DISPLAY=${DISPLAY}
    • WAYLAND_DISPLAY=${WAYLAND_DISPLAY}
    • XDG_RUNTIME_DIR=${XDG_RUNTIME_DIR}
    • PULSE_SERVER=${PULSE_SERVER} volumes:
    • /tmp/.X11-unix:/tmp/.X11-unix
    • /mnt/wslg:/mnt/wslg ```

4️⃣ Gazebo is Not Using GPU (Not Visible in nvidia-smi)

Issue

Gazebo is running, but nvidia-smi does not show it using the GPU.

Solution

  1. Run Gazebo with Vulkan explicitly enabled: 
    export MESA_D3D12_DEFAULT_ADAPTER_NAME=NVIDIA
gz sim -v 4
  2. Check nvidia-smi again in another terminal inside the container: 
    nvidia-smi

5️⃣ docker attach Freezes the Terminal

Issue

After running docker attach ros2-gpu-container, the terminal freezes when running commands.

Solution

  1. Use a new shell session instead: 
    docker exec -it ros2-gpu-container bash
  2. Or, detach from the session properly:
    • Press: Ctrl + P then Ctrl + Q

For more advanced issues, check Docker logs with:

docker logs ros2-gpu-container

✅ Need More Help?

This site is open source. Improve this page.