nvidia-runtime

算力节点，需要安装英伟达容器运行时

官方网址：

https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html#configuring-containerd-for-kubernetes

注册安装源（稳定版本）

curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

如果是体验学习，可使用实验版本

#!/bin/sh

sed -i -e '/experimental/ s/^#//g' /etc/apt/sources.list.d/nvidia-container-toolkit.list

更新安装 nvidia-container-toolkit

#!/bin/sh

sudo apt-get update

sudo apt-get install -y nvidia-container-toolkit

将 nvidia-runtime 配置到 containerd 中

安装完 nvidia-container-tookit 后，可以使用 nvidia-ctk 命令修改 containerd 配置

#!/bin/sh

sudo nvidia-ctk runtime configure --runtime=containerd

sudo systemctl restart containerd

上面的命令将 nvidia-runtime 加入到 containerd 的 runtime 列表中

如果需要，也可把 nvidia 设置为 containerd 的默认运行时

/etc/containerd/config.toml

    [plugins."io.containerd.grpc.v1.cri".containerd]
      default_runtime_name = "nvidia"

也可使用 shell 脚本更改默认运行时

#!/bin/sh

# 将默认 runtime 改为 nvidia
sed -i 's/default_runtime_name = "runc"/default_runtime_name = "nvidia"/g' /etc/containerd/config.toml