NVIDIA/nvidia-container-runtime

Fedora 35 unsupported, could centos8 work?

XuYuan114514 opened this issue · 3 comments

Since Fedora 35 is not in the list of supported distributions, I browsed through other similar issues. @elezar suggested to use the Centos repositories for installing packages.
So I run

curl -s -L https://nvidia.github.io/nvidia-container-runtime/centos8/nvidia-container-runtime.repo | \                                                        
  sudo tee /etc/yum.repos.d/nvidia-container-runtime.repo

and sudo yum install nvidia-container-runtime.
Then I restart my docker.
But in the end it still failed to run with [failed](docker: Error response from daemon: could not select device driver "" with capabilities: [[gpu]].)
So the code for my installation of centos8 version is wrong, or fedora 35 is not supported?

The centos8 packages should be compatible with fedora35 since the main limitation would be the minimum glibc version.

Note that your instructions have not added the NVIDIA Container Runtime to the docker /etc/docker/daemon.json file which is why You're seeing the error. We are in the process of updating our documentation to make things clearer across operating systems.

Please try the following instructions:

Add the repositories

distribution=centos8 \
                && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.repo | sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo

Note: distribution is being overridden to centos8

Install the NVIDIA Container Toolkit

sudo dnf clean expire-cache \
                && sudo dnf install -y nvidia-container-toolkit

Note: the nvidia-container-toolkit package is installed

Add the NVIDIA Container Runtime to the docker config

sudo nvidia-ctk runtime configure --runtime=docker

Restart the Docker daemon

sudo systemctl restart docker

If any of these steps give problems, please let us know.

elezar commented

@XuYuan114514 could you confirm that the instructions provided worked for you?