Vuklan not recognized with nvidia-toolkit 1.18, Debian 13 Trixie

[snowgoon88]

Hello,

I'm trying to use Vulkan through docker+nvidia-toolkit, but it fails because of faulty initialisation/configuration.

I have a manual fix below.
And it might be (somehow) related to issues #1472, #1021 and maybe #1517.

Problem: Vulkan not recognized ==============================================

In docker, the GPU is not see by Vulkan

[host]$ docker run -it --privileged --net=host --env="DISPLAY" --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" --gpus all --runtime=nvidia nvidia/cuda:12.4.1-cudnn-devel-ubuntu22.04 bash -c "apt update && apt install vulkan-tools -y && nvidia-smi && vulkaninfo --summary"

+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.163.01             Driver Version: 550.163.01     CUDA Version: 12.4     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 4060 Ti     On  |   00000000:07:00.0  On |                  N/A |
|  0%   34C    P8             15W /  165W |    1214MiB /  16380MiB |      7%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
+-----------------------------------------------------------------------------------------+
ERROR: [Loader Message] Code 0 : libnvidia-glsi.so.550.163.01: cannot open shared object file: No such file or directory
WARNING: [Loader Message] Code 0 : terminator_CreateInstance: Failed to CreateInstance in ICD 2.  Skipping ICD.
error: XDG_RUNTIME_DIR not set in the environment.
error: XDG_RUNTIME_DIR not set in the environment.
error: XDG_RUNTIME_DIR not set in the environment.
error: XDG_RUNTIME_DIR not set in the environment.
error: XDG_RUNTIME_DIR not set in the environment.
==========
VULKANINFO
==========

Vulkan Instance Version: 1.3.204


Instance Extensions: count = 20
-------------------------------
VK_EXT_acquire_drm_display             : extension revision 1
VK_EXT_acquire_xlib_display            : extension revision 1
VK_EXT_debug_report                    : extension revision 10
VK_EXT_debug_utils                     : extension revision 2
VK_EXT_direct_mode_display             : extension revision 1
VK_EXT_display_surface_counter         : extension revision 1
VK_EXT_swapchain_colorspace            : extension revision 4
VK_KHR_device_group_creation           : extension revision 1
VK_KHR_display                         : extension revision 23
VK_KHR_external_fence_capabilities     : extension revision 1
VK_KHR_external_memory_capabilities    : extension revision 1
VK_KHR_external_semaphore_capabilities : extension revision 1
VK_KHR_get_display_properties2         : extension revision 1
VK_KHR_get_physical_device_properties2 : extension revision 2
VK_KHR_get_surface_capabilities2       : extension revision 1
VK_KHR_surface                         : extension revision 25
VK_KHR_surface_protected_capabilities  : extension revision 1
VK_KHR_wayland_surface                 : extension revision 6
VK_KHR_xcb_surface                     : extension revision 6
VK_KHR_xlib_surface                    : extension revision 6

Instance Layers: count = 4
--------------------------
VK_LAYER_INTEL_nullhw       INTEL NULL HW                1.1.73   version 1
VK_LAYER_MESA_device_select Linux device selection layer 1.3.211  version 1
VK_LAYER_MESA_overlay       Mesa Overlay layer           1.3.211  version 1
VK_LAYER_NV_optimus         NVIDIA Optimus layer         1.3.277  version 1

Devices:
========
GPU0:
	apiVersion         = 4206847 (1.3.255)
	driverVersion      = 1 (0x0001)
	vendorID           = 0x10005
	deviceID           = 0x0000
	deviceType         = PHYSICAL_DEVICE_TYPE_CPU
	deviceName         = llvmpipe (LLVM 15.0.7, 256 bits)
	driverID           = DRIVER_ID_MESA_LLVMPIPE
	driverName         = llvmpipe
	driverInfo         = Mesa 23.2.1-1ubuntu3.1~22.04.3 (LLVM 15.0.7)
	conformanceVersion = 1.3.1.1
	deviceUUID         = 6d657361-3233-2e32-2e31-2d3175627500
	driverUUID         = 6c6c766d-7069-7065-5555-494400000000

Directly on the host machine:

[host]$ nvidia-smi 
Tue Jan  6 13:49:56 2026       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.163.01             Driver Version: 550.163.01     CUDA Version: 12.4     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 4060 Ti     Off |   00000000:07:00.0  On |                  N/A |
|  0%   34C    P8             15W /  165W |    1370MiB /  16380MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A      1534      G   /usr/lib/xorg/Xorg                            295MiB |
|    0   N/A  N/A      3495      G   /usr/lib/thunderbird/thunderbird              170MiB |
|    0   N/A  N/A    522193      G   firefox-esr                                   891MiB |
+-----------------------------------------------------------------------------------------+

[host]$ vulkaninfo --summary
==========
VULKANINFO
==========

Vulkan Instance Version: 1.4.309


Instance Extensions: count = 24
-------------------------------
VK_EXT_acquire_drm_display             : extension revision 1
VK_EXT_acquire_xlib_display            : extension revision 1
VK_EXT_debug_report                    : extension revision 10
VK_EXT_debug_utils                     : extension revision 2
VK_EXT_direct_mode_display             : extension revision 1
VK_EXT_display_surface_counter         : extension revision 1
VK_EXT_headless_surface                : extension revision 1
VK_EXT_surface_maintenance1            : extension revision 1
VK_EXT_swapchain_colorspace            : extension revision 5
VK_KHR_device_group_creation           : extension revision 1
VK_KHR_display                         : extension revision 23
VK_KHR_external_fence_capabilities     : extension revision 1
VK_KHR_external_memory_capabilities    : extension revision 1
VK_KHR_external_semaphore_capabilities : extension revision 1
VK_KHR_get_display_properties2         : extension revision 1
VK_KHR_get_physical_device_properties2 : extension revision 2
VK_KHR_get_surface_capabilities2       : extension revision 1
VK_KHR_portability_enumeration         : extension revision 1
VK_KHR_surface                         : extension revision 25
VK_KHR_surface_protected_capabilities  : extension revision 1
VK_KHR_wayland_surface                 : extension revision 6
VK_KHR_xcb_surface                     : extension revision 6
VK_KHR_xlib_surface                    : extension revision 6
VK_LUNARG_direct_driver_loading        : extension revision 1

Instance Layers: count = 4
--------------------------
VK_LAYER_INTEL_nullhw       INTEL NULL HW                1.1.73   version 1
VK_LAYER_MESA_device_select Linux device selection layer 1.4.303  version 1
VK_LAYER_MESA_overlay       Mesa Overlay layer           1.4.303  version 1
VK_LAYER_NV_optimus         NVIDIA Optimus layer         1.3.277  version 1

Devices:
========
GPU0:
	apiVersion         = 1.3.277
	driverVersion      = 550.163.1.0
	vendorID           = 0x10de
	deviceID           = 0x2805
	deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
	deviceName         = NVIDIA GeForce RTX 4060 Ti
	driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
	driverName         = NVIDIA
	driverInfo         = 550.163.01
	conformanceVersion = 1.3.7.2
	deviceUUID         = 895229b1-52bf-a326-6565-cbd4d20a56e3
	driverUUID         = 2485a78d-e39e-5aba-a0e9-b1139a1e6395
GPU1:
	apiVersion         = 1.4.305
	driverVersion      = 0.0.1
	vendorID           = 0x10005
	deviceID           = 0x0000
	deviceType         = PHYSICAL_DEVICE_TYPE_CPU
	deviceName         = llvmpipe (LLVM 19.1.7, 256 bits)
	driverID           = DRIVER_ID_MESA_LLVMPIPE
	driverName         = llvmpipe
	driverInfo         = Mesa 25.0.7-2 (LLVM 19.1.7)
	conformanceVersion = 1.3.1.1
	deviceUUID         = 6d657361-3235-2e30-2e37-2d3200000000
	driverUUID         = 6c6c766d-7069-7065-5555-494400000000

Context: ===========================================================
OS Debian 13 Trixie
nvidia driver: 550.163.01 installed with apt-get

[host]$ ls /usr/lib/x86_64-linux-gnu/libnvidia*   

/usr/lib/x86_64-linux-gnu/libnvidia-allocator.so.1          /usr/lib/x86_64-linux-gnu/libnvidia-glcore.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-cfg.so.1                /usr/lib/x86_64-linux-gnu/libnvidia-glsi.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-container-go.so.1       /usr/lib/x86_64-linux-gnu/libnvidia-glvkspirv.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-container-go.so.1.18.1  /usr/lib/x86_64-linux-gnu/libnvidia-gpucomp.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-container.so.1
/usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1
/usr/lib/x86_64-linux-gnu/libnvidia-container.so.1.18.1
/usr/lib/x86_64-linux-gnu/libnvidia-nvvm.so.4
/usr/lib/x86_64-linux-gnu/libnvidia-eglcore.so.550.163.01   /usr/lib/x86_64-linux-gnu/libnvidia-nvvm.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-egl-gbm.so.1
/usr/lib/x86_64-linux-gnu/libnvidia-opencl.so.1
/usr/lib/x86_64-linux-gnu/libnvidia-egl-gbm.so.1.1.2        /usr/lib/x86_64-linux-gnu/libnvidia-pkcs11-openssl3.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-egl-wayland.so.1        /usr/lib/x86_64-linux-gnu/libnvidia-ptxjitcompiler.so.1
/usr/lib/x86_64-linux-gnu/libnvidia-egl-wayland.so.1.1.18   /usr/lib/x86_64-linux-gnu/libnvidia-rtcore.so.550.163.01
/usr/lib/x86_64-linux-gnu/libnvidia-encode.so.1             /usr/lib/x86_64-linux-gnu/libnvidia-tls.so.550.163.01

[host]$ ls /usr/lib/x86_64-linux-gnu/nvidia/current 
libcuda.so                         libGLX_nvidia.so.0                 libnvidia-encode.so             libnvidia-opencl.so.550.163.01
libcuda.so.1                       libGLX_nvidia.so.550.163.01        libnvidia-encode.so.1           libnvidia-ptxjitcompiler.so.1
libcuda.so.550.163.01              libnvcuvid.so                      libnvidia-encode.so.550.163.01  libnvidia-ptxjitcompiler.so.550.163.01
libEGL_nvidia.so.0                 libnvcuvid.so.1                    libnvidia-ml.so                 libvdpau_nvidia.so.1
libEGL_nvidia.so.550.163.01        libnvcuvid.so.550.163.01           libnvidia-ml.so.1               libvdpau_nvidia.so.550.163.01
libGLESv1_CM_nvidia.so.1           libnvidia-allocator.so.1           libnvidia-ml.so.550.163.01      nvidia-drm_gbm.so
libGLESv1_CM_nvidia.so.550.163.01  libnvidia-allocator.so.550.163.01  libnvidia-nvvm.so.4
libGLESv2_nvidia.so.2              libnvidia-cfg.so.1                 libnvidia-nvvm.so.550.163.01
libGLESv2_nvidia.so.550.163.01     libnvidia-cfg.so.550.163.01        libnvidia-opencl.so.1


[host]$ docker --version
Docker version 29.1.3, build f52814d

[host]$ nvidia-container-cli --version 
cli-version: 1.18.1
lib-version: 1.18.1
build date: 2025-11-24T14:45+00:00
build revision: 889a3bb5408c195ed7897ba2cb8341c7d249672f
build compiler: x86_64-linux-gnu-gcc-7 7.5.0
build platform: x86_64
build flags: -D_GNU_SOURCE -D_FORTIFY_SOURCE=2 -DNDEBUG -std=gnu11 -O2 -g -fdata-sections -ffunction-sections -fplan9-extensions -fstack-protector -fno-strict-aliasing -fvisibility=hidden -Wall -Wextra -Wcast-align -Wpointer-arith -Wmissing-prototypes -Wnonnull -Wwrite-strings -Wlogical-op -Wformat=2 -Wmissing-format-attribute -Winit-self -Wshadow -Wstrict-prototypes -Wunreachable-code -Wconversion -Wsign-conversion -Wno-unknown-warning-option -Wno-format-extra-args -Wno-gnu-alignof-expression -Wl,-zrelro -Wl,-znow -Wl,-zdefs -Wl,--gc-sections

[host]$ less /etc/docker/daemon.json
{
    "runtimes": {
        "nvidia": {
            "args": [],
            "path": "nvidia-container-runtime"
        }
    }
}

Manual solution : because of configuration issues ============================================

In fact, I discovered that /etc/vulkan was not populated (still empty, even after restarting nvidia-cdi-refresh.service which is supposed to do the job) and that /etc/cdi/nvidia.yaml was missing some links (even after sudo nvidia-ctk cdi generate --output=/etc/cdi/nvidia.yaml

So I manually added the following symbolic links

[host]$ tree -l /etc/vulkan
.
├── explicit_layer.d
├── icd.d
│   └── nvidia_icd.json -> /usr/share/vulkan/icd.d/nvidia_icd.json
└── implicit_layer.d
    └── nvidia_layers.json -> /usr/share/vulkan/implicit_layer.d/nvidia_layers.json

And I added the following lines at the end of /etc/cdi/nvidia.yaml
(note that all these links are not in /usr/lib/x86_64-linux-gnu/nvidia/current but direclty in /usr/lib/x86_64-linux-gnu

== added to /etc/cdi/nvidia.yaml
        - hostPath: /usr/lib/x86_64-linux-gnu/libnvidia-glsi.so.550.163.01
          containerPath: /usr/lib/x86_64-linux-gnu/libnvidia-glsi.so.550.163.01
          options:
            - ro
            - nosuid
            - nodev
            - rbind
            - rprivate
        - hostPath: /usr/lib/x86_64-linux-gnu/libnvidia-tls.so.550.163.01
          containerPath: /usr/lib/x86_64-linux-gnu/libnvidia-tls.so.550.163.01
          options:
            - ro
            - nosuid
            - nodev
            - rbind
            - rprivate
        - hostPath: /usr/lib/x86_64-linux-gnu/libnvidia-glcore.so.550.163.01
          containerPath: /usr/lib/x86_64-linux-gnu/libnvidia-glcore.so.550.163.01
          options:
            - ro
            - nosuid
            - nodev
            - rbind
            - rprivate
        - hostPath: /usr/lib/x86_64-linux-gnu/libnvidia-gpucomp.so.550.163.01
          containerPath: /usr/lib/x86_64-linux-gnu/libnvidia-gpucomp.so.550.163.01
          options:
            - ro
            - nosuid
            - nodev
            - rbind
            - rprivate
        - hostPath: /usr/lib/x86_64-linux-gnu/libnvidia-glvkspirv.so.550.163.01
          containerPath: /usr/lib/x86_64-linux-gnu/libnvidia-glvkspirv.so.550.163.01
          options:
            - ro
            - nosuid
            - nodev
            - rbind
            - rprivate

Then I can use --env NVIDIA_VISIBLE_DEVICES=nvidia.com/gpu=all in running the docker and Vulkan is okay

[host]$ docker run -it --privileged --net=host --env="DISPLAY" --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" -e NVIDIA_VISIBLE_DEVICES=nvidia.com/gpu=all --runtime=nvidia ghcr.io/nvidia/k8s-samples:vulkan-86d426f7 vulkaninfo --summary

error: XDG_RUNTIME_DIR not set in the environment.
==========
VULKANINFO
==========

Vulkan Instance Version: 1.4.313


Instance Extensions: count = 23
-------------------------------
VK_EXT_acquire_drm_display             : extension revision 1
VK_EXT_acquire_xlib_display            : extension revision 1
VK_EXT_debug_report                    : extension revision 10
VK_EXT_debug_utils                     : extension revision 2
VK_EXT_direct_mode_display             : extension revision 1
VK_EXT_display_surface_counter         : extension revision 1
VK_EXT_surface_maintenance1            : extension revision 1
VK_EXT_swapchain_colorspace            : extension revision 4
VK_KHR_device_group_creation           : extension revision 1
VK_KHR_display                         : extension revision 23
VK_KHR_external_fence_capabilities     : extension revision 1
VK_KHR_external_memory_capabilities    : extension revision 1
VK_KHR_external_semaphore_capabilities : extension revision 1
VK_KHR_get_display_properties2         : extension revision 1
VK_KHR_get_physical_device_properties2 : extension revision 2
VK_KHR_get_surface_capabilities2       : extension revision 1
VK_KHR_portability_enumeration         : extension revision 1
VK_KHR_surface                         : extension revision 25
VK_KHR_surface_protected_capabilities  : extension revision 1
VK_KHR_wayland_surface                 : extension revision 6
VK_KHR_xcb_surface                     : extension revision 6
VK_KHR_xlib_surface                    : extension revision 6
VK_LUNARG_direct_driver_loading        : extension revision 1

Instance Layers: count = 10
---------------------------
VK_LAYER_KHRONOS_profiles         Khronos Profiles layer                                                                                            1.4.313  version 1
VK_LAYER_KHRONOS_shader_object    Khronos Shader object layer                                                                                       1.4.313  version 1
VK_LAYER_KHRONOS_synchronization2 Khronos Synchronization2 layer                                                                                    1.4.313  version 1
VK_LAYER_KHRONOS_validation       Khronos Validation Layer                                                                                          1.4.313  version 1
VK_LAYER_LUNARG_api_dump          LunarG API dump layer                                                                                             1.4.313  version 2
VK_LAYER_LUNARG_crash_diagnostic  Crash Diagnostic Layer is a crash/hang debugging tool that helps determines GPU progress in a Vulkan application. 1.4.313  version 1
VK_LAYER_LUNARG_gfxreconstruct    GFXReconstruct Capture Layer Version 1.0.5-unknown                                                                1.4.313  version 4194309
VK_LAYER_LUNARG_monitor           Execution Monitoring Layer                                                                                        1.4.313  version 1
VK_LAYER_LUNARG_screenshot        LunarG image capture layer                                                                                        1.4.313  version 1
VK_LAYER_NV_optimus               NVIDIA Optimus layer                                                                                              1.3.277  version 1

Devices:
========
GPU0:
	apiVersion         = 1.3.277
	driverVersion      = 550.163.1.0
	vendorID           = 0x10de
	deviceID           = 0x2805
	deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
	deviceName         = NVIDIA GeForce RTX 4060 Ti
	driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
	driverName         = NVIDIA
	driverInfo         = 550.163.01
	conformanceVersion = 1.3.7.2
	deviceUUID         = 895229b1-52bf-a326-6565-cbd4d20a56e3
	driverUUID         = 2485a78d-e39e-5aba-a0e9-b1139a1e6395

[elezar]

Thanks for the detailed investigation. It is unexpected from the point of view of the NVIDIA Container Toolkit that different driver components have different parent folders. We use the parent folder of libcuda.so.VERSION as the search path for driver libraries and this is why the additional libraries are missed.

For the vulkan ICDs, our detection logic in general can be improved there. I will look into what can be done to make this more robust. As a matter of interest. Are the ICD symlinks REQUIRED, or is the injection of the libraries sufficient?

Could you also please provide the target of the /usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1 symlink and confirm whether a libcuda.so.1 symlink also exists in /usr/lib/x86_64-linux-gnu/. (Here my assumption is that these point to the respective libraries in /usr/lib/x86_64-linux-gnu/current).

[snowgoon88]

Hello Elezar,
and thank you for being so reactive.

1. if I remove the ICD symlinks in /etc/vuklan, the container will not run with the --env NVIDIA_VISIBLE_DEVICE :

[host]$ docker run -it --privileged --net=host --env="DISPLAY" --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" -e NVIDIA_VISIBLE_DEVICES=nvidia.com/gpu=all --runtime=nvidia ghcr.io/nvidia/k8s-samples:vulkan-86d426f7 vulkaninfo --summary
docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: failed to fulfil mount request: open /etc/vulkan/icd.d/nvidia_icd.json: no such file or directory

And using the runtime approach (with --gpus=all, it fails with a different error message:

[host]$ docker run -it --privileged --net=host --env="DISPLAY" --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" --gpus=all --runtime=nvidia ghcr.io/nvidia/k8s-samples:vulkan-86d426f7 vulkaninfo --summary
ERROR: [Loader Message] Code 0 : libnvidia-glsi.so.550.163.01: cannot open shared object file: No such file or directory
ERROR: [Loader Message] Code 0 : loader_icd_scan: Failed loading library associated with ICD JSON libGLX_nvidia.so.0. Ignoring this JSON
ERROR: [Loader Message] Code 0 : vkCreateInstance: Found no drivers!
Cannot create Vulkan instance.
This problem is often caused by a faulty installation of the Vulkan driver or attempting to use a GPU that does not support Vulkan.
ERROR at ./vulkaninfo/./vulkaninfo.h:573:vkCreateInstance failed with ERROR_INCOMPATIBLE_DRIVER

2. the target of /usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1

[host]$ ls -l  /usr/lib/x86_64-linux-gnu/libnvidia*
...
lrwxrwxrwx 1 root root   60 Jul 21 23:28 /usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1 -> /etc/alternatives/nvidia--libnvidia-ml.so.1-x86_64-linux-gnu
...

# and
[host$ ls -l /etc/alternatives/nvidia--libnvidia-ml.so.1-x86_64-linux-gnu
lrwxrwxrwx 1 root root 58 Nov 21 16:35 /etc/alternatives/nvidia--libnvidia-ml.so.1-x86_64-linux-gnu -> /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-ml.so.1

3. for libcuda

[host]$  ls -l /usr/lib/x86_64-linux-gnu/libcuda*
lrwxrwxrwx 1 root root 53 Jul 21 23:28 /usr/lib/x86_64-linux-gnu/libcuda.so -> /etc/alternatives/nvidia--libcuda.so-x86_64-linux-gnu
lrwxrwxrwx 1 root root 55 Jul 21 23:28 /usr/lib/x86_64-linux-gnu/libcuda.so.1 -> /etc/alternatives/nvidia--libcuda.so.1-x86_64-linux-gnu

# and
[host]$ ls -l /etc/alternatives/nvidia--libcuda.so-x86_64-linux-gnu
lrwxrwxrwx 1 root root 51 Nov 21 16:35 /etc/alternatives/nvidia--libcuda.so-x86_64-linux-gnu -> /usr/lib/x86_64-linux-gnu/nvidia/current/libcuda.so

So, they do point to /usr/lib/x86_64-linux-gnu/current, but indirectly because of the alternatives way of choosing some libraries and application in debian.

Hope this helps,
Alain

[elezar]

Thanks @snowgoon88. Looking at the Debian package definitions for the drivers here: https://salsa.debian.org/nvidia-team/nvidia-graphics-drivers/-/tree/550?ref_type=heads it seems that not all driver libraries are installed to /usr/lib/x86_64-linux-gnu/current. I am following up internally to check what the expectations are with respect to this.

[miraculin1]

Hello elezar,

I can reproduce what looks like the same underlying issue on Debian 13 with the NVIDIA 550.163.01 driver.

In my case, the problem is not limited to Vulkan: glxinfo -B inside the container also fails to detect the NVIDIA device.

A useful extra point is that, inside the container:

ldd /usr/lib/x86_64-linux-gnu/nvidia/current/libGLX_nvidia.so.0

reports missing NVIDIA GL-related dependencies, including:

• libnvidia-glsi.so.550.163.01
• libnvidia-tls.so.550.163.01
• libnvidia-glcore.so.550.163.01

These files exist on the host under /usr/lib/x86_64-linux-gnu/.

As a temporary workaround, manually bind-mounting the missing libraries into the container makes glxinfo -B work again.

If useful, I can provide the exact docker run command and full ldd output.