Ubuntu18.04安装docker-ce、以及nvidia-docker

平台：Ubuntu 18.04.4 LTS
GPU：1080Ti 双卡

1. 首先如果你之前装过docker并且没有成功，那么先全部卸载掉和docker相关的所有安装过的包

sudo apt-get remove docker* --purge

2. 安装包以允许通过HTTPS使用存储库

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

3. 添加Docker的官方GPG密钥

curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
sudo apt-key fingerprint 0EBFCD88

输出如下：

pub   rsa4096 2017-02-22 [SCEA]
      9DC8 5822 9FC7 DD38 854A  E2D8 8D81 803C 0EBF CD88
uid           [ 未知 ] Docker Release (CE deb) <docker@docker.com>
sub   rsa4096 2017-02-22 [S]

4. 添加到源

即使您还想从边缘或测试存储库安装构建，您始终需要稳定的存储库。要添加边缘或测试存储库，请在以下命令中的单词stable之后添加单词edge或test（或两者）

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

5. 最后

sudo apt-get install -y nvidia-docker2

注意，到这就结束是因为docker-ce18.09.0已经自动安装 如果颠倒顺序就是：

sudo apt-get install docker-ce=18.09.0
sudo apt-get install -y nvidia-docker2

PS：这个时候如果提示E: 无法定位软件包 nvidia-docker2
那就先执行下面这一句话，之后再执行sudo apt-get install -y nvidia-docker2：

curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | \
  sudo apt-key add -
distribution="ubuntu18.04"
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | \
  sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update

6. 安装一个docker容器进行测试

sudo pkill -SIGHUP dockerd

sudo docker run --runtime=nvidia --rm nvidia/cuda:10.1-base nvidia-smi

这句话就是安装nvidia/cuda:10.1-base这个容器，之后在容器中执行nvidia-smi命令。
输出如下：

(base) xx@xxx:~$ sudo docker run --runtime=nvidia --rm nvidia/cuda:10.1-base nvidia-smi
Unable to find image 'nvidia/cuda:10.1-base' locally
10.1-base: Pulling from nvidia/cuda
7ddbc47eeb70: Pull complete
c1bbdc448b72: Pull complete
8c3b70e39044: Pull complete
45d437916d57: Pull complete
d8f1569ddae6: Pull complete
85386706b020: Pull complete
ee9b457b77d0: Pull complete
Digest: sha256:3cb86d1437161ef6998c4a681f2ca4150368946cc8e09c5e5178e3598110539f
Status: Downloaded newer image for nvidia/cuda:10.1-base
Thu Mar 12 09:49:37 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 430.50       Driver Version: 430.50       CUDA Version: 10.1     |
|-------------------------------+----------------------+----------------------+
| 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 GTX 108...  Off  | 00000000:17:00.0 Off |                  N/A |
| 31%   48C    P2   157W / 250W |  10891MiB / 11178MiB |     51%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX 108...  Off  | 00000000:65:00.0  On |                  N/A |
| 29%   27C    P8    10W / 250W |    449MiB / 11177MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
+-----------------------------------------------------------------------------+

PS：nvidia docker指定显卡启动
docker 启动参数上加 -e NVIDIA_VISIBLE_DEVICES=1(0代表第一块显卡，１代表第二块显卡，不加参数是全部启动)

sudo docker run -it  --runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=1 需要启动的docker名称 bash

给docker分配内存

REFERENCE

https://cloud.tencent.com/developer/article/1421365
https://blog.csdn.net/wuzhongli/article/details/86539433
https://www.jianshu.com/p/5bdcd4804456