-
Notifications
You must be signed in to change notification settings - Fork 227
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Can't run container using nvidia-docker - libnvidia-ml.so.1 #219
Comments
Hello again, can't anyone help me out? @elezar do you know how can I fix this issue? I believe that if I redirect the location of the file nvidia-docker is looking for to one of the locations the file is present in my PC it should this issue, However I don't know how to do it? Thanks. |
Hi, I am also experiencing the same issue. It is worth mentioning that I have Docker Desktop installed. The issue occurs when building an image using the sudo docker command, I can get the image to recognize the GPU and the nvidia-smi command inside docker works correctly. However, this does not work in a development environment if I want to use this image, as it is not seen by vscode. On the other hand, when I am trying to build the image using docker build and then running it, I get the error, mentioned by OP: From what I have seen, it may be an issue with Docker Desktop: Any way to solve this? Thank you very much. |
We do not support Docker Desktop. Please use docker-ce instead. |
Hi @elezar in my case the Nvidia-smi command doesn't even run. I am still waiting for some help since I created this issue and didn't get any! I believe my prints should help, if you could help me |
@pfcouto since you are using rootless docker, would using podman be an option for you? We have recently added support for generating CDI specifications using the Please see https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html#container-device-interface-cdi-support for information on generating CDI specifications and consuming them using podman. If podman is not an option, the NVIDIA Container Runtime can be configured in CDI mode to allow CDI specifications to be consumed by docker (also in rootless mode) until support is available. This is expected as an experimental feature in the Docker 25 release. |
For me, the reason for the error was a curropted installation of docker-ce, so I deleted docker completely and re-installed it using latest instruction on official documentation. |
you have to start and build the container with sudo in docker-ce |
On EndeavourOS (Arch) solved by removing Unfortunately, this means that I can't run anything that expects docker desktop... That's a really dumb incompatibility. |
1. Issue or feature description
Upon running the command
docker run --privileged --gpus all nvidia/cuda:11.0.3-base-ubuntu20.04 nvidia-smi
i get the error2. Steps to reproduce the issue
I installed nvidia through Fedora Docs, not Nvidia, so as an example
nvcc --version
outputs an error saying that it does not recognize nvcc command but in my host machine I can runnvidia-smi
The commands I used to install nvidia are the following:
And as visible in the following image I am able to run the command
nvidia-smi
in my host machineI followed this guide on how yo install nvidia-docker - - and did the following:
and upon running this docker command I get the error show in ### 1.
The thing is, I have the file that it says it is missing (check the following image), so maybe it is looking for it in a different directory?
3. Information to attach (optional if deemed irrelevant)
uname -a
:docker version
rpm -qa '*nvidia*'
nvidia-container-cli -V
Thanks for your help!
The text was updated successfully, but these errors were encountered: