Creating and Configuring /etc/docker/daemon.json to Resolve GPU Unavailability in Docker Container

[Shin-ichi-Takayama]

GPU was available in the running Docker container, but after a while it became unavailable.
Running the nvidia-smi command showed Failed to initialize NVML: Unknown Error.
I found the following solution.
NVIDIA/nvidia-docker#1671 (comment)

However, there was no /etc/docker/daemon.json file.
Can I create the /etc/docker/daemon.json file and what should I write in it to fix the problem?

Best regards.

[elezar]

The solution you linked does not mention the /etc/docker/daemon.json. Have you performed the other steps discussed in the issue?

[Shin-ichi-Takayama]

The link states that.

2.Explicitly disabling systemd cgroup management in Docker:.
Set the parameter
"exec-opts": ["native.cgroupdriver=cgroupfs"] in the /etc/docker/daemon.json file and restart docker.

No, I have not tried any other method.
Method No. 2 in the link is the easiest to implement.

[elezar]

Ah ... sorry. Missed that. How is docker configured on your system? It should be fine to create the file at /etc/docker/daemon.json with the contents:

{
    "exec-opts": ["native.cgroupdriver=cgroupfs"]
}

then.

[Shin-ichi-Takayama]

Thank you for your response. I will try that method.