Dockerr installation environment ubuntu16.04 + CUDA10.0 + cudnn = 7.0 + Anaconda + tensorflow-gpu = 1.13.1

Docker colleagues need to create for each server container, in order to develop, test. Close-up following the tutorial, make a simple record

1, the server environment

ubuntu, Nvidia RTX2080ti * 4, CUDA10.0

2, Docker installation

Docker divided into two versions: Docker EE Enterprise Edition, Docker CE Personal Edition free

Refer to the official document: https://docs.docker.com/install/linux/docker-ce/ubuntu/

Installation Docker:

a、卸载旧版本
$ sudo apt-get remove docker docker-engine docker.io containerd runc

b、更新apt包索引
$ sudo apt-get update

c、安装软件包以允许apt通过HTTPS使用存储库
$ sudo apt-get install \
    apt-transport-https \
    ca-certificates \
    curl \
    gnupg-agent \
    software-properties-common
停顿输入y

d、添加Docker的官方GPG秘钥
$ curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -

e、设置稳定的存储库
$ sudo add-apt-repository \
   "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
   $(lsb_release -cs) \
   stable"

f、再次更新apt软件包索引
sudo apt-get update

g、安装最新版本的Docker CE
sudo apt-get install docker-ce
停顿输入y

h、检测
docker --version
如果出现版本内容，那么安装成功

或者：
sudo docker run hello-world

i、创建docker组
$ sudo groupadd docker

j、将用户添加到该docker组
$ sudo usermod -aG docker $USER        # 当前用户
$ sudo usermod -aG docker gaofei        # 用户gaofei

Uninstall Docker (unloaded when it is used):

a、卸载Docker CE软件包
$ sudo apt-get purge docker-ce

b、主机上的图像，容器，卷或自定义配置文件不会自动删除。要删除所有图像，容器和卷：
$ sudo rm -rf /var/lib/docker

3, nvidia-docker installation

https://blog.csdn.net/chxw098/article/details/79741586

4, Docker installation image

5, create Dockerr container

Create a container command:

docker run -id --name=gaofei -v /home/docker/docker_data/gaofei_data/:/home/data/ -p 6666:6666 tfgpu1.13.1-ubuntu16.04-cuda10.0-v1.0:latest   /bin/bash

Reference documents:

https://www.cnblogs.com/program-meow/p/7994776.html   straightforward docker installation tutorial

https://docs.docker.com/install/linux/docker-ce/ubuntu/      official installation documentation docker

https://www.cnblogs.com/xhyan/p/6593075.html      into the docker container of four ways