Dockerfile 就是用来构建docker镜像的构建文件。命令脚本，先体验一下。
通过这个脚本可以生成镜像，这个镜像是一层一层的，脚本一个个的命令，每个命令是一层

  1. # 创建一个dockerfile 文件 名字可以随意，建议dockerfile
  2. # 文件中的内容 指令（大写）  参数
  3. FROM centos
  5. VOLUME ["/volume01","/volume02"]
  7. CMD echo "--------end--------"
  8. CMD /bin/bash
  9. # 这里的每条命令就是镜像的一层
  11. # 注意这里的 VOLUME ["/volume01","/volume02"] 通过VOLUME指令指定了两个挂载点 /volume01 和 /volume02
  12. # 不能写成VOLUME ["volume01","volume02"] 要加 / 否则在启动时会报错
  13. [root@localhost docker-test-volume]# docker run -it 4a64df3f0173 /bin/bash
  14. docker: Error response from daemon: OCI runtime create failed: invalid mount {Destination:volume01 Type:bind Source:/var/lib/docker/volumes/3918fef01cbe6da3036cdf9d2a3d461b3fe365de11d9b20858a5631aa7db9282/_data Options:[rbind]}: mount destination volume01 not absolute: unknown.
  16. # 生成镜像
  17. [root@localhost docker-test-volume]# docker build -f /home/docker-test-volume/dockerfile1 -t ceshi/centos:1.0 .

image.png
启动自己生成的镜像
image.png
我们通过docker inspect 查看通过该dockerfile创建的镜像生成的容器，可以看到如下信息
image.png
由此推断出使用的是匿名挂载方式，既挂载在宿主机默认路径的随机目录下/var/lib/docker/volumes/xxx/_data
测试一下文件是否同步

  1. # 容器内创建一个文件
  2. [root@localhost docker-test-volume]# docker run -it 2e245ff80b78  /bin/bash
  3. [root@c27e4c77cce9 /]# cd volume01
  4. [root@c27e4c77cce9 volume01]# touch container.txt
  5. [root@c27e4c77cce9 volume01]# ls
  6. container.txt
  8. # 容器外查看是否文件同步
  9. [root@localhost home]# cd /var/lib/docker/volumes/45d97f75807b1cf296163ee9a189a422b1e39188578b9f1e594a2cf2537327b1/_data
  10. [root@localhost _data]# ls
  11. container.txt
  13. # 文件同步成功

这种方式使用的十分多，因为我们通常会构建自己的镜像！
假设构建镜像时候没有挂载卷，要手动镜像挂载 -v 卷名：容器内路径！

最佳实践：数据卷容器

如果多个容器需要共享数据（如持久化数据库、配置文件或者数据文件等），可以考虑创建一个特定的数据卷容器，该容器有1个或多个卷。其它容器通过--volumes-from 来共享这个数据卷容器的卷。

启动第一个容器docker01

  1. [root@localhost docker-test-volume]# docker run -it --name docker01 ceshi/centos:1.0
  2. [root@14d2256f90b7 /]# ls -l
  3. total 0
  4. lrwxrwxrwx.   1 root root   7 Nov  3  2020 bin -> usr/bin
  5. drwxr-xr-x.   5 root root 360 Jun  9 04:00 dev
  6. drwxr-xr-x.   1 root root  66 Jun  9 04:00 etc
  7. drwxr-xr-x.   2 root root   6 Nov  3  2020 home
  8. lrwxrwxrwx.   1 root root   7 Nov  3  2020 lib -> usr/lib
  9. lrwxrwxrwx.   1 root root   9 Nov  3  2020 lib64 -> usr/lib64
  10. drwx------.   2 root root   6 Dec  4  2020 lost+found
  11. drwxr-xr-x.   2 root root   6 Nov  3  2020 media
  12. drwxr-xr-x.   2 root root   6 Nov  3  2020 mnt
  13. drwxr-xr-x.   2 root root   6 Nov  3  2020 opt
  14. dr-xr-xr-x. 100 root root   0 Jun  9 04:00 proc
  15. dr-xr-x---.   2 root root 162 Dec  4  2020 root
  16. drwxr-xr-x.  11 root root 163 Dec  4  2020 run
  17. lrwxrwxrwx.   1 root root   8 Nov  3  2020 sbin -> usr/sbin
  18. drwxr-xr-x.   2 root root   6 Nov  3  2020 srv
  19. dr-xr-xr-x.  13 root root   0 Jun  8 13:19 sys
  20. drwxrwxrwt.   7 root root 145 Dec  4  2020 tmp
  21. drwxr-xr-x.  12 root root 144 Dec  4  2020 usr
  22. drwxr-xr-x.  20 root root 262 Dec  4  2020 var
  23. drwxr-xr-x.   2 root root   6 Jun  9 04:00 volume01
  24. drwxr-xr-x.   2 root root   6 Jun  9 04:00 volume02
  25. [root@14d2256f90b7 /]#

启动第二个容器docker02，并通过—volumes-from 来共享数据

  1. [root@localhost docker-test-volume]# docker run -it --name docker02 --volumes-from docker01 ceshi/centos:1.0
  2. [root@5cfb1e6eb1ff /]# ls -l
  3. total 0
  4. lrwxrwxrwx.   1 root root   7 Nov  3  2020 bin -> usr/bin
  5. drwxr-xr-x.   5 root root 360 Jun  9 04:13 dev
  6. drwxr-xr-x.   1 root root  66 Jun  9 04:13 etc
  7. drwxr-xr-x.   2 root root   6 Nov  3  2020 home
  8. lrwxrwxrwx.   1 root root   7 Nov  3  2020 lib -> usr/lib
  9. lrwxrwxrwx.   1 root root   9 Nov  3  2020 lib64 -> usr/lib64
  10. drwx------.   2 root root   6 Dec  4  2020 lost+found
  11. drwxr-xr-x.   2 root root   6 Nov  3  2020 media
  12. drwxr-xr-x.   2 root root   6 Nov  3  2020 mnt
  13. drwxr-xr-x.   2 root root   6 Nov  3  2020 opt
  14. dr-xr-xr-x. 103 root root   0 Jun  9 04:13 proc
  15. dr-xr-x---.   2 root root 162 Dec  4  2020 root
  16. drwxr-xr-x.  11 root root 163 Dec  4  2020 run
  17. lrwxrwxrwx.   1 root root   8 Nov  3  2020 sbin -> usr/sbin
  18. drwxr-xr-x.   2 root root   6 Nov  3  2020 srv
  19. dr-xr-xr-x.  13 root root   0 Jun  8 13:19 sys
  20. drwxrwxrwt.   7 root root 145 Dec  4  2020 tmp
  21. drwxr-xr-x.  12 root root 144 Dec  4  2020 usr
  22. drwxr-xr-x.  20 root root 262 Dec  4  2020 var
  23. drwxr-xr-x.   2 root root   6 Jun  9 04:11 volume01
  24. drwxr-xr-x.   2 root root   6 Jun  9 04:11 volume02

在docker01中创建的数据同步到了docker02中
image.png
再创建一个docker03也共享数据到docker01

  1. [root@localhost docker-test-volume]# docker run -it --name docker03 --volumes-from docker01 ceshi/centos:1.0
  2. [root@51d0c3deb4cd /]# ls -l
  3. total 0
  4. lrwxrwxrwx.   1 root root   7 Nov  3  2020 bin -> usr/bin
  5. drwxr-xr-x.   5 root root 360 Jun  9 04:23 dev
  6. drwxr-xr-x.   1 root root  66 Jun  9 04:23 etc
  7. drwxr-xr-x.   2 root root   6 Nov  3  2020 home
  8. lrwxrwxrwx.   1 root root   7 Nov  3  2020 lib -> usr/lib
  9. lrwxrwxrwx.   1 root root   9 Nov  3  2020 lib64 -> usr/lib64
  10. drwx------.   2 root root   6 Dec  4  2020 lost+found
  11. drwxr-xr-x.   2 root root   6 Nov  3  2020 media
  12. drwxr-xr-x.   2 root root   6 Nov  3  2020 mnt
  13. drwxr-xr-x.   2 root root   6 Nov  3  2020 opt
  14. dr-xr-xr-x. 106 root root   0 Jun  9 04:23 proc
  15. dr-xr-x---.   2 root root 162 Dec  4  2020 root
  16. drwxr-xr-x.  11 root root 163 Dec  4  2020 run
  17. lrwxrwxrwx.   1 root root   8 Nov  3  2020 sbin -> usr/sbin
  18. drwxr-xr-x.   2 root root   6 Nov  3  2020 srv
  19. dr-xr-xr-x.  13 root root   0 Jun  8 13:19 sys
  20. drwxrwxrwt.   7 root root 145 Dec  4  2020 tmp
  21. drwxr-xr-x.  12 root root 144 Dec  4  2020 usr
  22. drwxr-xr-x.  20 root root 262 Dec  4  2020 var
  23. drwxr-xr-x.   2 root root  22 Jun  9 04:16 volume01
  24. drwxr-xr-x.   2 root root   6 Jun  9 04:11 volume02
  25. [root@51d0c3deb4cd /]# cd volume01
  26. [root@51d0c3deb4cd volume01]# ls
  27. docker01
  28. [root@51d0c3deb4cd volume01]# touch docker03
  29. [root@51d0c3deb4cd volume01]# ls
  30. docker01  docker03
  31. [root@51d0c3deb4cd volume01]# 
  33. # docker01中也可以看到docker03创建的文件
  34. [root@14d2256f90b7 volume01]# ls
  35. docker01  docker03

测试将docker01彻底删除，查看docker02上的文件

  1. # 删除docker01
  2. [root@localhost _data]# docker ps -a
  3. CONTAINER ID   IMAGE              COMMAND                  CREATED          STATUS                            PORTS     NAMES
  4. 51d0c3deb4cd   ceshi/centos:1.0   "/bin/sh -c /bin/bash"   8 minutes ago    Exited (127) About a minute ago             docker03
  5. 5cfb1e6eb1ff   ceshi/centos:1.0   "/bin/sh -c /bin/bash"   17 minutes ago   Up 17 minutes                               docker02
  6. 14d2256f90b7   ceshi/centos:1.0   "/bin/sh -c /bin/bash"   20 minutes ago   Exited (0) 38 seconds ago                   docker01
  7. [root@localhost _data]# docker rm -f 14d2256f90b7
  8. 14d2256f90b7
  10. # 查看docker02中的数据
  11. [root@localhost docker-test-volume]# docker attach 5cfb1e6eb1ff
  12. [root@5cfb1e6eb1ff volume01]# ls
  13. docker01  docker03
  14. [root@5cfb1e6eb1ff volume01]# 
  15. # 在docker02中创建文件docker02
  16. [root@localhost docker-test-volume]# docker attach 5cfb1e6eb1ff
  17. [root@5cfb1e6eb1ff volume01]# ls
  18. docker01  docker03
  19. [root@5cfb1e6eb1ff volume01]# touch docker02
  20. [root@5cfb1e6eb1ff volume01]# ls
  21. docker01  docker02  docker03
  22. #在docker03中查看
  23. [root@localhost ~]# docker attach 51d0c3deb4cd
  24. [root@51d0c3deb4cd /]# cd volume01
  25. [root@51d0c3deb4cd volume01]# ls
  26. docker01  docker03
  27. [root@51d0c3deb4cd volume01]# ls
  28. docker01  docker02  docker03
  30. # 可以看到数据好好的，依旧可以进行同步，为什么？
  31. # 因为容器的卷本质上对应主机上的目录，所以这个数据卷容器不需要启动，甚至删除也没关系。
  1. # 验证 查看docker02 docker03中的文件挂载路径
  2. # 发现docker02 docker03 都映射在了宿主机的同一个路径下
  3. [root@localhost _data]# docker inspect docker02
  4.         ...
  5.         "Mounts": [
  6.             {
  7.                 "Type": "volume",
  8.                 "Name": "2630c69c3583e0df3e6cf1af00e8a858d58463d7259f76aaaa788dcdc463b63b",
  9.                 "Source": "/var/lib/docker/volumes/2630c69c3583e0df3e6cf1af00e8a858d58463d7259f76aaaa788dcdc463b63b/_data",
  10.                 "Destination": "/volume01",
  11.                 "Driver": "local",
  12.                 "Mode": "",
  13.                 "RW": true,
  14.                 "Propagation": ""
  15.             },
  16.             {
  17.                 "Type": "volume",
  18.                 "Name": "adbe88a9686e56ee290ceb5eaaed37541e45cebdb7bdb3daad8af97d28cbf7fd",
  19.                 "Source": "/var/lib/docker/volumes/adbe88a9686e56ee290ceb5eaaed37541e45cebdb7bdb3daad8af97d28cbf7fd/_data",
  20.                 "Destination": "/volume02",
  21.                 "Driver": "local",
  22.                 "Mode": "",
  23.                 "RW": true,
  24.                 "Propagation": ""
  25.             }
  26.         ],
  27.         ...
  28. [root@localhost _data]# docker inspect docker03
  29.                     ...
  30.           "Mounts": [
  31.               {
  32.                   "Type": "volume",
  33.                   "Name": "2630c69c3583e0df3e6cf1af00e8a858d58463d7259f76aaaa788dcdc463b63b",
  34.                   "Source": "/var/lib/docker/volumes/2630c69c3583e0df3e6cf1af00e8a858d58463d7259f76aaaa788dcdc463b63b/_data",
  35.                   "Destination": "/volume01",
  36.                   "Driver": "local",
  37.                   "Mode": "",
  38.                   "RW": true,
  39.                   "Propagation": ""
  40.               },
  41.               {
  42.                   "Type": "volume",
  43.                   "Name": "adbe88a9686e56ee290ceb5eaaed37541e45cebdb7bdb3daad8af97d28cbf7fd",
  44.                   "Source": "/var/lib/docker/volumes/adbe88a9686e56ee290ceb5eaaed37541e45cebdb7bdb3daad8af97d28cbf7fd/_data",
  45.                   "Destination": "/volume02",
  46.                   "Driver": "local",
  47.                   "Mode": "",
  48.                   "RW": true,
  49.                   "Propagation": ""
  50.               }
  51.           ],
  52.           ...

结论：
容器之间的配置信息的传递，数据卷容器的生命周期一直持续到没有容器使用为止。
但是一旦你持久化到了本地，这个时候，本地的数据是不会删除的！