docker总结

常用命令

打印image列表

docker images

打印container列表

docker container ls
docker container ls --all

拉取官方镜像

docker pull ubuntu:16.04

拉取自己的镜像

docker login(登录账号)
docker image pull timleezhong/helloworld

删除镜像

docker image rm <REPOSITORY:TAG>或者IMAGE ID>
docker rmi <REPOSITORY:TAG>或者IMAGE ID>

保存镜像为tar包

docker image save -o ubuntu.tar.gz ubuntu:16.04
docker save -o ubuntu.tar.gz ubuntu:16.04

从tar包导入镜像

docker load -i <tar file>

docker load -i ubuntu.tar.gz

推送自己的镜像到仓库

docker login
docker image push timleezhong/helloworld

创建容器

docker run -d -it -p 8022:22 --name ubuntu ubuntu:16.04
docker container run --publish 8000:8080 --detach --name ubuntu ubuntu:16.04

启动容器

docker start ubuntu

在正在运行的容器中运行一条命令

docker exec -it ubuntu bash
docker exec -it ubuntu python
docker exec -it ubuntu ls

删除容器

docker container rm --force ubuntu

提交容器到镜像

docker commit CONTAINER REPOSITORY:TAG

---

从Dockerfile构建镜像

• docker build -t caffe:gpu standalone/gpu 【从standalone/gpu目录下的Dockerfile构建镜像caffe:gpu】
• docker image build -t bulletinboard:1.0 .

临时使用container

• docker run -it ubuntu bash 【try something more ambitious热切】

入门教程

https://docs.docker.com/get-started/part2/

用Dockerfile定义一个容器

# Use the official image as a parent image
FROM node:current-slim

# Set the working directory
WORKDIR /usr/src/app

# Copy the file from your host to your current location
COPY package.json .

# Run the command inside your image filesystem
RUN npm install

# Inform Docker that the container is listening on the specified port at runtime.
EXPOSE 8080

# Run the specified command within the container.
CMD [ "npm", "start" ]

# Copy the rest of your app's source code from your host to your image filesystem.
COPY . .

建立并测试您的镜像

进入Dockerfile所在目录，然后执行：
docker image build -t bulletinboard:1.0 .

bulletinboard:1.0表示REPOSITORY:TAG

将镜像作为容器运行

docker container run --publish 8000:8080 --detach --name bb bulletinboard:1.0

删除：

docker container rm --force bb
该–force选项将删除正在运行的容器。

发布镜像

准备工作：创建docker账号

docker login
docker image push timleezhong/helloworld

docker image pull timleezhong/helloworld

拉取官方镜像

docker pull ubuntu:16.04

安装

https://docs.docker.com/install/linux/docker-ce/ubuntu/
https://github.com/NVIDIA/nvidia-docker

阉割版安装
apt install docker.io
systemctl start docker
docker pull ubuntu
docker save -o ubuntu.tar.gz ubuntu:latest

配置

允许非root用户使用

 sudo usermod -aG docker your-user

测试

sudo docker run hello-world

Usage

#### Test nvidia-smi with the latest official CUDA image
$ docker run --gpus all nvidia/cuda:9.0-base nvidia-smi

# Start a GPU enabled container on two GPUs
$ docker run --gpus 2 nvidia/cuda:9.0-base nvidia-smi

# Starting a GPU enabled container on specific GPUs
$ docker run --gpus '"device=1,2"' nvidia/cuda:9.0-base nvidia-smi
$ docker run --gpus '"device=UUID-ABCDEF,1"' nvidia/cuda:9.0-base nvidia-smi

# Specifying a capability (graphics, compute, ...) for my container
# Note this is rarely if ever used this way
$ docker run --gpus all,capabilities=utility nvidia/cuda:9.0-base nvidia-smi

离线安装

docker pull nvidia/7.5-cudnn5-runtime-ubuntu14.04

docker save -o 7.5-cudnn5-runtime-ubuntu14.04.tar.gz nvidia/cuda:7.5-cudnn5-runtime-ubuntu14.04

docker load < 7.5-cudnn5-runtime-ubuntu14.04.tar.gz

docker run -d  --shm-size 60G --name cudnn75 --restart=always -v ~/torch0.2_env:/workdir --gpus all -p 1022:22 nvidia/cuda:7.5-cudnn4-devel-ubuntu14.04

docker run -d  --shm-size 60G --name cudnn75 --restart=always -v ~/torch0.2_env:/workdir --gpus all -p 1022:22 nvidia/cuda:7.5-cudnn4-devel-ubuntu14.04

docker exec -it cuda-75 /bin/bash