Docker GPU not showing up

[MrCrunshy]

Hello, I'm completely new to this. I've set up Glances via Docker, but my RTX 3050 isn't showing up.

My Jellyfin instance is running inside the Docker server and can access it without any problems.

Unfortunately, I can't find the cause of the problem.

My system (Ubuntu) is completely up to date.

Here's an excerpt from my Compose file:

version: '3'

services:
  monitoring:
    image: nicolargo/glances:latest-full
    pid: host
    network_mode: host
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
      # Uncomment the below line if you want glances to display host OS detail instead of container's
      # - /etc/os-release:/etc/os-release:ro
    environment:
      - "GLANCES_OPT=-w"
    # For nvidia GPUs
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

To Reproduce
Steps to reproduce the behavior:

1. Start Glances with Docker
2. Enter the Webgui
3. See error

Screenshots

Environement (please complete the following information)

• Operating System (lsb_release -a or OS name/version): To be completed with result of: lsb_release -a

No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 24.04.3 LTS
Release: 24.04
Codename: noble

If you need any further information, please let me know.
Greets
Crunshy

[nicolargo]

Hi @MrCrunshy

Please run the Glances in debug mode (GLANCES_OPT=-w -d) and paste/bin the glances.log file located in the Glances container (most of the time in /tmp/glances-root.log).

Thanks

[MrCrunshy]

And here is the log file :)
glances-root.log

[nicolargo]

@MrCrunshy found this log message:

2025-12-28 21:57:00,710 -- DEBUG -- NVML Shared Library (libnvidia-ml.so.1) not Found, Nvidia GPU plugin is disabled

Under the wood, Glances use https://pypi.org/project/nvidia-ml-py/ to grab GPU stats.

Please can you try this inside your Glances countainer:

python
>>> from pynvml import *
>>> nvmlInit()
>>> print(f"Driver Version: {nvmlSystemGetDriverVersion()}")
Driver Version: 11.515.48
>>> deviceCount = nvmlDeviceGetCount()
>>> for i in range(deviceCount):
...     handle = nvmlDeviceGetHandleByIndex(i)
...     print(f"Device {i} : {nvmlDeviceGetName(handle)}")
...
Device 0 : Tesla K40c
>>> nvmlShutdown()

Thanks !

[MrCrunshy]

I get the first error during nvmlInit().

/app # python
Python 3.12.12 (main, Oct 11 2025, 15:09:54) [GCC 14.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from pynvml import *
>>> nvmlInit()
Traceback (most recent call last):
  File "/venv/lib/python3.12/site-packages/pynvml.py", line 2824, in _LoadNvmlLibrary
    nvmlLib = CDLL("libnvidia-ml.so.1")
              ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.12/ctypes/__init__.py", line 379, in __init__
    self._handle = _dlopen(self._name, mode)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^
OSError: Error loading shared library libnvidia-ml.so.1: No such file or directory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/venv/lib/python3.12/site-packages/pynvml.py", line 2796, in nvmlInit
    nvmlInitWithFlags(0)
  File "/venv/lib/python3.12/site-packages/pynvml.py", line 2779, in nvmlInitWithFlags
    _LoadNvmlLibrary()
  File "/venv/lib/python3.12/site-packages/pynvml.py", line 2826, in _LoadNvmlLibrary
    _nvmlCheckReturn(NVML_ERROR_LIBRARY_NOT_FOUND)
  File "/venv/lib/python3.12/site-packages/pynvml.py", line 1061, in _nvmlCheckReturn
    raise NVMLError(ret)
pynvml.NVMLError_LibraryNotFound: NVML Shared Library Not Found

[mrMightyM]

I have the same issue with a RTX 3060, docker compose is the same as above, the GPU works in other containers. I enabled debugging, and tried the python code in the container and get the same error after the nvmlInit() call.

[nicolargo]

@mrMightyM can you try to install the Nvidia Container Toolkit on your host. It should expose your GPU to the container.

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

And try to run your Glances container again.

[mrMightyM]

@nicolargo I already do have that installed. I use the GPU from other containers with no issue - Immich, Audiomuse and others all work fine. Glances just does not see it.

From the host:

root@odin:~# dpkg --list *toolkit*
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name                            Version      Architecture Description
+++-===============================-============-============-============================================
ii  cuda-toolkit                    13.1.0-1     amd64        CUDA Toolkit meta-package
ii  cuda-toolkit-13-1               13.1.0-1     amd64        CUDA Toolkit 13.1 meta-package
ii  cuda-toolkit-13-1-config-common 13.1.80-1    all          Common config package for CUDA Toolkit 13.1.
ii  cuda-toolkit-13-config-common   13.1.80-1    all          Common config package for CUDA Toolkit 13.
ii  cuda-toolkit-config-common      13.1.80-1    all          Common config package for CUDA Toolkit.
ii  nvidia-container-toolkit        1.18.1-1     amd64        NVIDIA Container toolkit
ii  nvidia-container-toolkit-base   1.18.1-1     amd64        NVIDIA Container Toolkit Base

root@odin:~# nvidia-smi
Fri Jan  2 09:08:22 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 590.44.01              Driver Version: 590.44.01      CUDA Version: 13.1     |
+-----------------------------------------+------------------------+----------------------+
| 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 3060        On  |   00000000:01:00.0 Off |                  N/A |
|  0%   41C    P8             16W /  170W |     396MiB /  12288MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A            2170      G   /usr/lib/xorg/Xorg                        4MiB |
|    0   N/A  N/A          621504      C   python3                                 268MiB |
|    0   N/A  N/A         1018006      C   python3                                 108MiB |
+-----------------------------------------------------------------------------------------+

From another container (Immich Machine Learning):

root@odin:~# docker exec -it immich_machine_learning /bin/bash
root@6a041a9378d6:/usr/src# nvidia-smi
Fri Jan  2 17:09:57 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 590.44.01              Driver Version: 590.44.01      CUDA Version: 13.1     |
+-----------------------------------------+------------------------+----------------------+
| 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 3060        On  |   00000000:01:00.0 Off |                  N/A |
|  0%   41C    P8             16W /  170W |     396MiB /  12288MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

From Glances container:

root@odin:~# docker exec -it glances /bin/sh
/app # ls /usr/bin/nvidia-*
/usr/bin/nvidia-cuda-mps-control  /usr/bin/nvidia-debugdump         /usr/bin/nvidia-smi
/usr/bin/nvidia-cuda-mps-server   /usr/bin/nvidia-persistenced
/app # /usr/bin/nvidia-smi
/bin/sh: /usr/bin/nvidia-smi: not found

Let me know if you need any other information.

[nicolargo]

Can you try to run Glances docker with the following command line (after stopping your Glances Docker compose):

docker run \
  --gpus all \
  --name glances \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -p 61208-61209:61208-61209 \
  --pid host \
  nicolargo/glances:latest-full

and test again.

[mrMightyM]

Different error:

root@odin:/home/docker/glances# docker compose down glances
[+] down 1/1
root@odin:/home/docker/glances# docker run --gpus all --name glances -v /var/run/docker.sock:/var/run/docker.sock:ro -p 61208-61209:61208-61209 --pid host nicolargo/glances:latest-full
Cannot init the curses library (setupterm: could not find terminal)

I've never seen that before.

[nicolargo]

Add -it:

docker run -it \
  --gpus all \
  --name glances \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -p 61208-61209:61208-61209 \
  --pid host \
  nicolargo/glances:latest-full

[mrMightyM]

Ok, that brought glances up in the terminal. I do have this down the right side:

/lib/firmware/nvidia/_  65.5G   125G
/lib/firmware/nvidia/_  65.5G   125G
/usr/bin/nvidia-cuda-_  65.5G   125G

with a bunch of other similar entries below that

Should I see GPU usage anywhere?

[nicolargo]

Yes you should see the GPU plugin on top (between CPU and memory plugins)...

Try to run it with -d (debug):

docker run -it \
  -e GLANCES_OPT="-d" \
  --gpus all \
  --name glances \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -p 61208-61209:61208-61209 \
  --pid host \
  nicolargo/glances:latest-full

Then enter in the running container (with docker exec) and have a look to the /tmp/glances-root.log file) for line about GPU (paste bin the file if possible).

[mrMightyM]

Ok, there is no GPU between CPU and memory.

New to pastebin, I think this link will work?

https://pastebin.com/Pp2ZXHHL

The logs are growing exponentially so I've copied an pasted a portion of it, if you need more let me know

[nicolargo]

I need the startup sequence. Please clean the log. Restart Glances and paste bin the log file.

[mrMightyM]

Sorry, lets try this link

https://pastebin.com/5ChAtjNv

I did see this near the beginning

2026-01-02 19:45:26,765 -- DEBUG -- NVML Shared Library (libnvidia-ml.so.1) not Found, Nvidia GPU plugin is disabled

If you need more of the log let me know.

[nicolargo]

Last chance, try with the Ubuntu based image (ubuntu latest full):

docker run -it \
  -e GLANCES_OPT="-d" \
  --gpus all \
  --name glances \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -p 61208-61209:61208-61209 \
  --pid host \
  nicolargo/glances:ubuntu-latest-full

[mrMightyM]

Okay, I tried that, and it works great! I guess I need to stick with the ubuntu image then? Is there much of a difference between it and the Alpine verion in regards to resource usage?

If there is anything else you want to try let me know. I feel like it should work with either image but I'm not sure how exactly it works

[MrCrunshy]

Last chance, try with the Ubuntu based image (ubuntu latest full):

docker run -it \
  -e GLANCES_OPT="-d" \
  --gpus all \
  --name glances \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -p 61208-61209:61208-61209 \
  --pid host \
  nicolargo/glances:ubuntu-latest-full

I tried it myself, but unfortunately nothing changed :(

[nicolargo]

Okay, I tried that, and it works great! I guess I need to stick with the ubuntu image then? Is there much of a difference between it and the Alpine verion in regards to resource usage?

If there is anything else you want to try let me know. I feel like it should work with either image but I'm not sure how exactly it works

Nope, the only difference is the image size (bigger for Ubuntu). I will add a note in the documentation to explain that only ubuntu based images are compliant with Glances GPU plugin. Thanks for your time.

[nicolargo]

Last chance, try with the Ubuntu based image (ubuntu latest full):

docker run -it \
  -e GLANCES_OPT="-d" \
  --gpus all \
  --name glances \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -p 61208-61209:61208-61209 \
  --pid host \
  nicolargo/glances:ubuntu-latest-full

I tried it myself, but unfortunately nothing changed :(

Is https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html installed on your host ?

[MrCrunshy]

Here, I've checked again, everything's there:

daniel@srv-docker:~$ export NVIDIA_CONTAINER_TOOLKIT_VERSION=1.18.1-1
  sudo apt-get install -y \
      nvidia-container-toolkit=${NVIDIA_CONTAINER_TOOLKIT_VERSION} \
      nvidia-container-toolkit-base=${NVIDIA_CONTAINER_TOOLKIT_VERSION} \
      libnvidia-container-tools=${NVIDIA_CONTAINER_TOOLKIT_VERSION} \
      libnvidia-container1=${NVIDIA_CONTAINER_TOOLKIT_VERSION}
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
nvidia-container-toolkit is already the newest version (1.18.1-1).
nvidia-container-toolkit set to manually installed.
nvidia-container-toolkit-base is already the newest version (1.18.1-1).
nvidia-container-toolkit-base set to manually installed.
libnvidia-container-tools is already the newest version (1.18.1-1).
libnvidia-container-tools set to manually installed.
libnvidia-container1 is already the newest version (1.18.1-1).
libnvidia-container1 set to manually installed.
0 upgraded, 0 newly installed, 0 to remove and 1 not upgraded.