Docker for GPU安装使用

安装Docker

• 卸载旧版本的Docker

sudo apt-get remove docker docker-engine docker-ce docker.io

• 跟新apt-get索引

sudo apt-get update

• 安装以下包以使apt可以通过HTTPS使用存储库（repository）

sudo apt-get install -y apt-transport-https ca-certificates curl software-properties-common

• 添加Docker官方GPG密钥

curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -

• 使用下面的命令来设置stable存储库：

sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"

• 更新apt索引

sudo apt-get update

• 安装

sudo apt-get install -y docker-ce

• 验证是否启动服务

systemctl status docker

• 若未启动，则启动dockerr

sudo systemctl start docker

• 打开hello word

sudo docker run hello-world

安装Nvidia Container Runtime for Ubantu

当使用Docker需要用到GPU时，出现错误：

docker: Error response from daemon: linux runtime spec devices: could not select device driver “” with capabilities: [[gpu]].

解决方案 安装nvidia-container-runtime

参考：Nvidia Container Toolkit installtion Guide

• 检测显卡驱动

lspci -vv | grep -i nvidia

• 设置stable仓库以及GPG密钥

distribution=$(. /etc/os-release;echo $ID$VERSION_ID)  

curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add  

curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list Note

curl -s -L https://nvidia.github.io/nvidia-container-runtime/experimental/$distribution/nvidia-container-runtime.list | sudo tee /etc/apt/sources.list.d/nvidia-container-runtime.list

注意： 最近这个repository被Q了，所以没法获取这个apt-get索引


在curl和apt-get后面附加上proxy可以解决这个问题

#ss
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
    
curl -x https://10.8.xxx.xxx:1080 -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add    

curl -x https://10.8.xxx.xxx:1080 -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list Note  

curl -x https://10.8.xxx.xxx:1080 -s -L https://nvidia.github.io/nvidia-container-runtime/experimental/$distribution/nvidia-container-runtime.list | sudo tee /etc/apt/sources.list.d/nvidia-container-runtime.list

• 更新索引

sudo apt-get update

#proxy
sudo apt-get -o Acquire::http::proxy="socks5h://10.8.xxx.xxx:1080/" update

• 安装

sudo apt-get install nvidia-container-runtime

#proxy
sudo apt-get -o Acquire::http::proxy="socks5h://10.8.xxx.xxx:1080/" install nvidia-container-runtime

• 查看

which  nvidia-container-runtime-hook

• 验证：运行利用GPU的容器

docker run -it --rm --gpus all ubuntu nvidia-smi

• 列出GPU设备

docker run -it --rm --gpus all ubuntu nvidia-smi -L