# UE Vulkan error in NVIDIA Docker

**URL:** <https://forums.unrealengine.com/t/ue-vulkan-error-in-nvidia-docker/2669674>\
**Category:** Platform & Builds\
**Tags:** Linux, linux-build, unreal-engine, question\
**Created:** [October 29, 2025, 12:28pm UTC](https://forums.unrealengine.com/t/ue-vulkan-error-in-nvidia-docker/2669674 "2025-10-29T12:28:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![sumitcasanova](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@sumitcasanova](https://forums.unrealengine.com/u/sumitcasanova)\
**Post date:** [October 29, 2025, 12:28pm UTC](https://forums.unrealengine.com/t/ue-vulkan-error-in-nvidia-docker/2669674/1 "2025-10-29T12:28:40Z")

</div>

I’m running an Unreal Engine 5.6 packaged Linux build inside a Docker container based on `nvidia/cuda:12.4.1-runtime-ubuntu22.04`. When launching the app, I get this error:

```auto
LogVulkanRHI: Error: vpCreateInstance failed, VkResult=-9 (VK_ERROR_INCOMPATIBLE_DRIVER)

```

I’ve installed `libvulkan1` and `vulkan-tools` inside the container and I’m launching with `xvfb-run`. However, Vulkan fails to initialize.

Dockerfile snippet:

```auto
FROM nvidia/cuda:12.4.1-runtime-ubuntu22.04
RUN apt-get update && apt-get install -y libvulkan1 vulkan-tools xvfb
ENV NVIDIA_VISIBLE_DEVICES=all
ENV NVIDIA_DRIVER_CAPABILITIES=all
CMD ["xvfb-run", "-a", "./AI_Assistant.sh"]

```

Run command:

```auto
docker run -it --gpus all --runtime=nvidia unreal-vulkan

```

The host machine has NVIDIA drivers and Vulkan working correctly (`vulkaninfo` runs fine outside Docker). I’m trying to understand why Vulkan isn’t initializing inside the container or how to make Unreal detect the correct Vulkan ICD.