把应用和运行环境一起打包到容器，然后用容器运行的技术。

可以理解为把本地写的程序通过 Docker 打包成一个“安装包”，再上传到像应用商城一样的镜像仓库，别人直接拉取这个镜像并运行，就能使用你的程序。

运行链路

本地代码 → Dockerfile → 构建镜像 → 上传镜像仓库 → 别人拉取镜像 → 启动容器运行

Docker常用命令

# view local images
docker images 

# pull image
docker pull image-name

# delete image
docker rmi image-id

# build image
# -f find the Dockerfile in the current directory
# -t display to label the image 
# version indicates which version of this image 
# . indicates build context which is the current directory
docker build -f Dockerfile --build-arg build-arg -t image-name:version .

# view running containers
docker ps

# view all containers
docker ps -a

# statt container
docker run -d \
-p  local-port:container-port \
--name container-name \
--network net-name \
-e container-env-varible(KEY=value) \
-v local-directory:container-directory \
image-name:version   

# stop container
docker stop container-name

# start the stopped container
docker start container-name

# restart container
docker restart container-name

# delete container
docker rm container-name

# force delete running container
docker rm -f container-name

# enter container
docker exec -it container-name /bin/bash
# if some container have not bash
docker exec -it container-name /bin/sh

# view container detailed information
docker inspect container-name

# view the usage of containers
docker stats

# view logs
docker logs container-name

# continuous tracking of logs
docker logs -f container-name

# view the last 100 lines of the log
docker logs --tail 100 container-name

# If the container is deleted, the internal data may be lost.
# so it is necessary to mount the data for preservation. 
# view data volume
docker volume ls

# delete data volume
docker volume rm volume-name

# view network
docker network ls

# create network
docker network create net-name

# delete all stopped containers
docker container prune

# delete dangling images
docker image prune

# clear unnecessary data
# -a indicates clear unnecessary image
docker system prune -a

Docker Compose

一个配置文件，统一管理多个容器，就是原来一串docker run命令统一写到compose.yaml里

# order compose
docker compose up -d

# stop and delete these container
docker compose down

# view service status
docker compose ps

# view logs
docker compose logs -f

# rebuild and start
docker compose up -d --build

比如说你现在要构建一个后端项目，需要用到MySQL和Redis，compose.yaml里面就可以这样写：

# indicate what services are there, which means what containers
services:
  app:
    # based on the Dockerfile in the directory
    build: .
    container_name: myapp
    ports:
      - "8080:8080"
    # startup sequence dependency
    depends_on:
      - mysql
      - redis
    networks:
      - mynet

  mysql:
    # indicate to use pre-build image directly
    image: mysql:8.0
    container_name: mysql
    environment:
      MYSQL_ROOT_PASSWORD: 123456
      MYSQL_DATABASE: testdb
    ports:
      - "3306:3306"
    volumes:
      - mysql_data:/var/lib/mysql
    networks:
      - mynet

  redis:
    image: redis:7
    container_name: redis
    ports:
      - "6379:6379"
    networks:
      - mynet

networks:
  mynet:

volumes:
  mysql_data:

Docker Compose 会自动给这一组服务建一个内部网络，放在同一个 compose.yml 里的服务，默认都接到这个网络上，所以它们彼此能直接通信

Dockerfile

构建镜像的说明书，告诉Docker基于什么基础镜像，需要哪些文件，需要安装哪些依赖，容器启动执行什么命令。

# the begging of this image
FROM 

# set work dir, like the cd in liunx
WORKDIR 

# take local dir copy to image
COPY

# copy-plus contain automatically extract compressed packages, support remote URLs
ADD

# execute commands when building image
RUN

# execute commands by default when container start
CMD

# main execute commands by default when container start
ENTRYPOINT

# java -jar app.jar
ENTRYPOINT ["java", "-jar"]
CMD ["app.jar"]

# set environment var
ENV

# declare the port in the internal container
EXPOSE

RUN 用于镜像构建阶段，通常执行安装依赖、编译构建等操作；

CMD 用于容器启动阶段，指定容器默认启动命令。

数据持久化

解决问题：容器删除了，数据怎么处理

why？

容器一般是临时的，其内部文件系统内数据，随着容器的删除也删除了。像数据库这类文件、上传日志等不能只放容器内部，必须挂载到容器外面保存。

bind mount

直接挂载宿主目录机

-v /local/mysql:/var/lib/mysql

左侧是电脑上目录，右侧是容器内MySQL存数据位置。主要是你管路径，可以通过cd进去看。更适合开发环境使用。

volume

挂载在Docker自己管理的存储空间

-v mysql:/var/lib/mysql

左侧mysql是一个Docker volume，Docker自己维护管理，更适合生产使用，迁移更方便。