三、Linux 安装Docker-（GPU）Nvidia

Docker安装 Nvidia-container-runtime 安装

如果需要安装支持GPU的Docker容器，则选择安装最新的docker安装版本19.03+。
1、卸载Docker和Nvidia-container-runtime。

yum remove -y docker-ce docker-ce-cli containerd
yum remove -y nvidia-container-runtime*  libnvidia-container*

2、执行如下命令，备份并移除daemon.json文件。

mkdir /etc/docker
vi /etc/docker/daemon.json

输入如下内容：

{
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "path": "/usr/bin/nvidia-container-runtime",
            "runtimeArgs": []
        }
    },
    "exec-opts": ["native.cgroupdriver=systemd"],
    "log-driver": "json-file",
    "log-opts": {
        "max-size": "100m",
        "max-file": "10"
    },
    "bip": "169.254.123.1/24",
    "oom-score-adjust": -1000,
    "storage-driver": "overlay2",
    "storage-opts":["overlay2.override_kernel_check=true"],
    "live-restore": true
}```
备份该配置：

```bash
mv /etc/docker/daemon.json /tmp/daemon.json

如果需要指定位置安装，新增参数graph即可，如下：

3、执行以下命令，安装Docker。
在待升级Docker版本的节点上下载Docker安装包。

VERSION=19.03.5 
URL=http://aliacs-k8s-cn-beijing.oss-cn-beijing.aliyuncs.com/public/pkg/docker/docker-${VERSION}.tar.gz 
curl -ssL $URL -o /tmp/docker-${VERSION}.tar.gz  
cd /tmp
tar -xf docker-${VERSION}.tar.gz
cd /tmp/pkg/docker/${VERSION}/rpm
yum localinstall -y $(ls .)

4、执行以下命令，在节点上安装Nvidia-container-runtime。

cd /tmp
yum install -y unzip
wget http://kubeflow.oss-cn-beijing.aliyuncs.com/nvidia.zip
unzip nvidia.zip
yum -y -q --nogpgcheck localinstall /tmp/nvidia/*

5、执行以下命令，配置daemon.json。
将上述的daemon.json覆盖/etc/docker/daemon.json，使原有配置生效。

mv /tmp/daemon.json  /etc/docker/daemon.json 

6、执行以下命令，重启Docker。

service docker start
systemctl enable docker

7、测试docker可正常使用nvidia

docker run --gpus all nvidia/cuda:10.2-base nvidia-smi

显示如下内容，则安装成功

[root@bogon tmp]# docker run --gpus all nvidia/cuda:10.2-base nvidia-smi
Unable to find image 'nvidia/cuda:10.2-base' locally
10.2-base: Pulling from nvidia/cuda
f08d8e2a3ba1: Pull complete 
3baa9cb2483b: Pull complete 
94e5ff4c0b15: Pull complete 
1860925334f9: Pull complete 
4e133546ace1: Pull complete 
db5ee31a93b2: Pull complete 
34fd2c745ff0: Pull complete 
Digest: sha256:4918ce9269396c77da6b45ee97e227fb56f37093c498a2a0bbebfda785a3b69e
Status: Downloaded newer image for nvidia/cuda:10.2-base
Thu Aug 20 05:01:57 2020       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.33.01    Driver Version: 440.33.01    CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:04:00.0 Off |                  N/A |
| 20%   33C    P0    11W / 250W |      0MiB / 11019MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

8、简单daemon.json配置如下：

{
  "registry-mirrors": [
    "https://registry.docker-cn.com",
    "http://hub-mirror.c.163.com",
    "https://docker.mirrors.ustc.edu.cn"
  ], 
  "insecure-registries": ["harbor.test.com","registry.cn-shenzhen.aliyuncs.com"], 
  "max-concurrent-downloads": 10
}

参考：

1、nvidia-docker离线安装包下载

http://mirror.cs.uchicago.edu/nvidia-docker/nvidia-container-runtime/stable/ubuntu16.04/amd64/
需要安装的包需要按照系统和需求自行选择。
参考：八、服务器【Ubuntu】GPU-TeslaP100部署中使用