Docker Deployment

RealityScan Linux is nearly production-ready and distributed as .deb and .rpm packages using CodeWeavers’ custom Wine runtime. Containerization enables consistent deployments across diverse Linux environments and is recommended for cloud automation, CI, and GPU-accelerated server workflows.

Why Docker?

• Native installers depend on specific kernel/glibc versions, causing compatibility issues across distros.

• Local installation introduces many dependencies that may conflict with system packages.

• Containers offer isolation, reproducibility, and simplified cloud deployment with GPU passthrough.

GPU Installation

NVIDIA drivers must be installed on the host system.

During OS installation, select:

Install third-party software for graphics and Wi-Fi hardware, and additional media formats

If not installed, use:

• NVIDIA driver installation guide

• Or install via https://www.nvidia.com/en-gb/drivers/

• For Ubuntu, consider using the NVIDIA PPA for the latest drivers.

Test GPU Detection

nvidia-smi

A correct setup should show the GPU table and CUDA version. 

Avoid nouveau drivers — Vulkan will not work and Wine GPU acceleration requires Vulkan.

Vulkan Installation

Wine requires Vulkan for GPU utilization.

Ubuntu

sudo apt install -y vulkan-tools vulkan-utils libvulkan1

Rocky, RHEL, Fedora

Enable Enterprise Toolset first:

sudo dnf install -y epel-release
sudo dnf install -y https://www.elrepo.org/elrepo-release-9.el9.elrepo.noarch.rpm

Then install Vulkan:

sudo dnf install -y vulkan vulkan-tools vulkan-loader

Test Vulkan

vulkaninfo --summary

This should show available Vulkan GPUs.

Docker Installation

Install Docker using: https://docs.docker.com/engine/install/

NVIDIA Container Toolkit

Required for GPU passthrough:  https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html

After installation, GPU passthrough works via:

--gpus=all

Dockerfile for RealityScan

Create a Dockerfile alongside the RealityScan .deb installer.

FROM ubuntu:22.04
ENV DEBIAN_FRONTEND=noninteractive
RUN apt-get update && \
    apt-get install -y --no-install-recommends \
      curl wget bzip2 xz-utils x11-apps ca-certificates \
      libgtk-3-0 libdbus-glib-1-2 libxt6 \
      libx11-xcb1 libxcb-shm0 libxcb-dri3-0 \
      libvulkan1 vulkan-tools \
      libxcomposite1 libasound2 && \
    rm -rf /var/lib/apt/lists/*

Build it:

docker build -t ubuntu-realityscan:v2.2.0.119039 .

Enabling Docker GPU Access

Enable Docker at startup:

sudo systemctl enable --now docker

Allow user to run Docker without sudo:

sudo usermod -aG docker $USER
newgrp docker

Allow X11 GUI Forwarding

xhost +local:docker

Running the Container

Same Distro (Host + Container)

docker run -it --gpus all \
-e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix \
-v ~/Downloads:/mnt/host \
-v /etc/vulkan/icd.d:/usr/share/vulkan/icd.d:ro \
-e NVIDIA_DRIVER_CAPABILITIES=all \
--device /dev/dri \
ubuntu-realityscan

Inside the container:

• nvidia-smi must work

• vulkaninfo --summary must match host GPU

Different Distros (Host ≠ Container)

You must manually map NVIDIA .so libraries. Example for Ubuntu host → Ubuntu 22.04 container:

docker run -it --gpus all \
-e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix \
-v /usr/share/vulkan/icd.d:/usr/share/vulkan/icd.d:ro \
-v /lib/x86_64-linux-gnu/libGLX_nvidia.so.0:/usr/lib64/libGLX_nvidia.so.0 \
-v /home/<user>/Downloads:/mnt/host \
-e NVIDIA_DRIVER_CAPABILITIES=all --device /dev/dri \
ubuntu-realityscan

Missing Vulkan files trigger:

Cannot create Vulkan instance.

Add missing nvidia .so files until Vulkan initializes.

Inside the Container

Run RealityScan:

/opt/realityscan/bin/realityscan

Check GPU availability under Settings → GPUs to use.

Test CUDA via:

/opt/realityscan/bin/wine CudaDeviceQuery.exe

Common Issues

• Some dialogs appear as black rectangles — press Enter.

• Epic login may need browser installation inside container.

• After exiting, restart with:

  Command Line

  docker start -ai <container>

  Copy full snippet(1 line long)