Python的docker-py模块

Posted 2022-05-19 礁之

tags:

篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了Python的docker-py模块相关的知识，希望对你有一定的参考价值。

文章目录

一、docker的配置

[root@centos-1 ~]# vim /etc/systemd/system/docker.service   #修改docker的系统启动文件  
——————————————————————————
 在'12'行添加'-H tcp://0.0.0.0:2375'，表示启动时占用端口,从而能够远程连接
 添加'-H unix'这段表示允许本地，不加重启后会使用docker命令会报错
 还有一个是tls可以使用ssl证书进行连接，但是需要指定证书位置
——————————————————————————
  1 [Unit]
  2 Description=Docker Application Container Engine
  3 Documentation=https://docs.docker.com
  4 After=network-online.target firewalld.service
  5 Wants=network-online.target
  6 
  7 [Service]
  8 Type=notify
  9 # the default is not to use systemd for cgroups because the delegate issues still
 10 # exists and systemd currently does not support the cgroup feature set required
 11 # for containers run by docker
 12 ExecStart=/usr/bin/dockerd  -H unix:///var/run/docker.sock -H tcp://0.0.0.0:2375  
 13 ExecReload=/bin/kill -s HUP $MAINPID
 14 # Having non-zero Limit*s causes performance problems due to accounting overhead
 15 # in the kernel. We recommend using cgroups to do container-local accounting.
 16 LimitNOFILE=infinity
 17 LimitNPROC=infinity
 18 LimitCORE=infinity
 19 # Uncomment TasksMax if your systemd version supports it.
 20 # Only systemd 226 and above support this version.
 21 #TasksMax=infinity
 22 TimeoutStartSec=0
 23 # set delegate yes so that systemd does not reset the cgroups of docker containers
 24 Delegate=yes
 25 # kill only the docker process, not all processes in the cgroup
 26 KillMode=process
 27 # restart the docker process if it exits prematurely
 28 Restart=on-failure
 29 StartLimitBurst=3
 30 StartLimitInterval=60s
 31 
 32 [Install]
 33 WantedBy=multi-user.target
#保存退出
[root@centos-1 ~]# systemctl daemon-reload
[root@centos-1 ~]# systemctl restart docker
[root@centos-1 ~]# netstat -ntpl | grep 2375
tcp6       0      0 :::2375                 :::*                    LISTEN      1313/dockerd   

###注意：
（1）如果开启防火墙则需要配置允许端口通过
（2）阿里云的ECS上部署的Docker，还需在安全组规则添加入方向、tcp、2375端口的开通

二、Python的docker模块

官网地址：Docker SDK 官网
docker-py 是一个第三方库，所以在使用之前需要自行安装。推荐使用 PIP 进行安装，命令如下：

- 先下载docker-py
pip install docker-py

- 安装好之后，进入交互模式查看版本
PS D:\\工作\\work> python
Python 3.10.4 (tags/v3.10.4:9d38120, Mar 23 2022, 23:13:41) [MSC v.1929 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import docker 
>>> docker.version_info
(1, 10, 6)
>>> docker.version
'1.10.6'

由于主要操作时针对docker服务器的，所以编写的代码基本上都算是客户端，在docker-py中最重要的操作基本上都在docker.client模块中，此模块提供了对docker引擎的相关操作

（1）建立连接

docker服务器对外提供的时web接口，即接收http消息并且返回http响应，但这里并不直接向docker服务发送http消息，而是使用客户端的高层接口函数来解决这个问题
使用接口函数相会更加直观、方便，并且和docker命令行客户端的用法类似
在对docker服务器进行操作之前，需要先建立连接，可以使用docker.from.env()接口函数来建立连接,大多数情况下不需要传入参数就可以正常建立连接，例如：
```
client = docker.from_env()
```
还有一种连接的方式是构造一个client对象，该对象的初始化函数可以接收下面的参数：
```
client = docker.Client(base_url='tcp://10.10.30.69:2375',version='1.21',timeout=5)  #远程连接
```
1. base_url：连接哪个 Docker 服务器，默认值是本机的 Docker 服务器。
2. version：使用的协议版本。
3. timeout：连接超时时间，单位为秒。如果在指定的时间内没有连接成功，则表示连接失败，返回值是 None。
4. tls：是否使用 TLS 作为传输层，默认值是 False，如果要使用的话，还需要指定ssl文件。

下面来看一个python如何与docker建立连接：

#在centos环境下
[root@centos-1 ~]# python3 -m venv tutorial-env  #创建虚拟环境目录
[root@centos-1 ~]# source tutorial-env/bin/activate  #进入虚拟环境
(tutorial-env) [root@centos-1 ~]# pip install docker-py  #下载docker，前面出现tutorial-env，表示进入虚拟环境
(tutorial-env) [root@centos-1 ~]# python3
Python 3.9.9 (main, May 13 2022, 15:23:56) 
[GCC 4.8.5 20150623 (Red Hat 4.8.5-16)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import docker
>>> docker.version
'1.10.6'
>>> client = docker.Client()
>>> if client is None:
...     print('connection error')
... else:
...     print('connection ok')
... 
connection ok     #连接成功
>>> client.base_url
'http+docker://localunixsocket'
>>> client.timeout
60
>>> ver_info = client.version()
>>> ver_info
'Platform': 'Name': '', 'Components': ['Name': 'Engine', 'Version': '18.06.1-ce', 'Details': 'ApiVersion': '1.38', 'Arch': 'amd64', 'BuildTime': '2018-08-21T17:28:38.000000000+00:00', 'Experimental': 'false', 'GitCommit': 'e68fc7a', 'GoVersion': 'go1.10.3', 'KernelVersion': '3.10.0-693.el7.x86_64', 'MinAPIVersion': '1.12', 'Os': 'linux'], 'Version': '18.06.1-ce', 'ApiVersion': '1.38', 'MinAPIVersion': '1.12', 'GitCommit': 'e68fc7a', 'GoVersion': 'go1.10.3', 'Os': 'linux', 'Arch': 'amd64', 'KernelVersion': '3.10.0-693.el7.x86_64', 'BuildTime': '2018-08-21T17:28:38.000000000+00:00'
>>> ver_info['Version']
'18.06.1-ce'
>>> ver_info['ApiVersion']
'1.38'
>>> ver_info['KernelVersion']
'3.10.0-693.el7.x86_64'
>>> ver_info['BuildTime']
'2018-08-21T17:28:38.000000000+00:00'

（2）对docker镜像的操作

docker的镜像就是软件的安装包，对其操作主要有下载、删除、推送等，对于到docker命令就是pull、push和rmi
如果想要下载某个镜像，默认的命令是docker pull imagename，例如：
```
docker pull alpine:3.9
```

使用python实现的话，可以这样写：

import docker
client = docker.from_env()
imgs1 = client.images()                    # 列出所有本地的镜像
client.pull("nginx")                    # 下载所有的alpine镜像，不要执行这个，这个会全部下载
imgs2 = client.images()                    # 再次查看所有的本地镜像
for img in imgs2:
    if img not in imgs1:
        print(img)

上面的代码会把alpine相关镜像全部下载，最好不要执行，我们可以加标签来指定下载的镜像：

import docker
client = docker.from_env()
imgs1 = client.images()                    # 列出所有本地的镜像
client.pull("nginx:1.20")                    # 下载所有的alpine镜像，不要执行这个，这个会全部下载
imgs2 = client.images()                    # 再次查看所有的本地镜像
for img in imgs2:
    if img not in imgs1:
        print(img)

还可以看到下载的过程

import docker
client = docker.from_env()
out_stream = client.pull("nginx:1.20", stream=True)
for line in out_stream:
    print(line)
    
    
#输出信息：
b'"status":"Downloading","progressDetail":"current":24861068,"total":25334886,"progress":"[=================================================\\\\u003e ]  24.86MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Downloading","progressDetail":"current":25123212,"total":25334886,"progress":"[=================================================\\\\u003e ]  25.12MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Verifying Checksum","progressDetail":,"id":"43bc2143131d"\\r\\n'
b'"status":"Download complete","progressDetail":,"id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":262144,"total":25334886,"progress":"[\\\\u003e                                                
  ]  262.1kB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":2097152,"total":25334886,"progress":"[====\\\\u003e                                           
   ]  2.097MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":3670016,"total":25334886,"progress":"[=======\\\\u003e                                        
   ]   3.67MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":5505024,"total":25334886,"progress":"[==========\\\\u003e                                     
   ]  5.505MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":7864320,"total":25334886,"progress":"[===============\\\\u003e                                
   ]  7.864MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":10223616,"total":25334886,"progress":"[====================\\\\u003e                          
    ]  10.22MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":13369344,"total":25334886,"progress":"[==========================\\\\u003e                    
    ]  13.37MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":15728640,"total":25334886,"progress":"[===============================\\\\u003e               
    ]  15.73MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":18087936,"total":25334886,"progress":"[===================================\\\\u003e           
    ]  18.09MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":20447232,"total":25334886,"progress":"[========================================\\\\u003e          ]  20.45MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":22806528,"total":25334886,"progress":"[=============================================\\\\u003e     ]  22.81MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":24379392,"total":25334886,"progress":"[================================================\\\\u003e  ]  24.38MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":25334886,"total":25334886,"progress":"[==================================================\\\\u003e]  25.33MB/25.33MB","id":"43bc2143131d"\\r\\n'
b'"status":"Pull complete","progressDetail":,"id":"43bc2143131d"\\r\\n'
b'"status":"Extracting","progressDetail":"current":601,"total":601,"progress":"[==================================================\\\\u003e]     601B/601B","id":"45e2f4bfb543"\\r\\n'
b'"status":"Extracting","progressDetail":"current":601,"total":601,"progress":"[==================================================\\\\u003e]     601B/601B","id":"45e2f4bfb543"\\r\\n'
b'"status":"Pull complete","progressDetail":,"id":"45e2f4bfb543"\\r\\n'
b'"status":"Extracting","progressDetail":"current":894,"total":894,"progress":"[==================================================\\\\u003e]     894B/894B","id":"18bc137763b4"\\r\\n'
b'"status":"Extracting","progressDetail":"current":894,"total":894,"progress":"[==================================================\\\\u003e]     894B/894B","id":"18bc137763b4"\\r\\n'
b'"status":"Pull complete","progressDetail":,"id":"18bc137763b4"\\r\\n'
b'"status":"Extracting","progressDetail":"current":664,"total":664,"progress":"[==================================================\\\\u003e]     664B/664B","id":"504f3151a203"\\r\\n'
b'"status":"Extracting","progressDetail":"current":664,"total":664,"progress":"[==================================================\\\\u003e]     664B/664B","id":"504f3151a203"\\r\\n'
b'"status":"Pull complete","progressDetail":,"id":"504f3151a203"\\r\\n'
b'"status":"Extracting","progressDetail":"current":1393,"total":1393,"progress":"[==================================================\\\\u003e]  1.393kB/1.393kB","id":"520efdb29b04"\\r\\n'
b'"status":"Extracting","progressDetail":"current":1393,"total":1393,"progress":"[==================================================\\\\u003e]  1.393kB/1.393kB","id":"520efdb29b04"\\r\\n'
b'"status":"Pull complete","progressDetail":,"id":"520efdb29b04"\\r\\n'
b'"status":"Digest: sha256:ec8e997ca71be5046070b2db6043fb7100e1681b91bb2fcb49d551b7fc9b2ba4"\\r\\n'
b'"status":"Status: Downloaded newer image for nginx:1.20"\\r\\n'

上面说了查看镜像和下载镜像，现在来看删除镜像：

import docker                                # 这个必须有，引入Docker包
client = docker.from_env()                    # 和Docker服务器建立连接
imgs = client.images()                        # 得到Docker镜像列表
for img in imgs:                            # 遍历所有的镜像
        # 一个镜像可能有几个tag，只要有一个符合要求即可
    for tag in img['RepoTags']:
        if tag == "nginx:1.20":                # tag是我们要找的
            client.remove_image(img['Id'])    # 删除挑选出来的镜像
            
            
 #执行脚本，之后可以使用docker images进行查看

有时，可能镜像已经创建容器，并且在运行中，那么上面的代码就会报错，可以使用强制删除的方式：

import docker                            # 必须要引入的Docker包
client = docker.from_env()                # 和Docker服务器建立连接   
imgs = client.images()                    # 得到所有的本地镜像
for img in imgs:
    for tag in img['RepoTags']:
        if tag == "nginx:1.20":             # 有我们要找的标签
            client.remove_image(img['Id'], force=True)    # 添加force=True进行强制删除

如果想要删除全部镜像：

import docker                         
client = docker.from_env()                # 和Docker服务器建立连接
imgs = client.images()                    # 得到所有的本地镜像
for img in imgs:                        # 依次遍历镜像文件
    client.remove_image(img['Id'])        # 直接删除镜像文件，不在向上面那样做if判断

还可以进行推送镜像，例如：

import docker
client = docker.from_env()
# 指定用户名、密码和服务器
login_ret = client.login(username="docker",     # 用户名
                    password="123.com",                     # 密码
                    registry="repo.docker.com")        # 服务器信息
print("login_ret:", login_ret)
# 将指定的镜像推送到服务器上
push_rsp = client.push("test:test", stream=False)
print("push_rsp:", push_rsp)

（3）使用Dockerfile文件创建镜像

如果想要使用Dockerfile创建镜像，可以这样使用：

——vim Dockerfile
    
FROM nginx:v1.0
CMD ["/bin/sh"]
#保存退出

——vim test.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import docker
client = docker.from_env()
# 注意，一定要以二进制方式打开
fd = open("./Dockerfile", "rb")
result = client.build(fileobj=fd, rm=True, tag='pydocker/demo')
fd.close()
for line in result:
    print(line)
#保存退出

——python3 test.py
#输出结果：
b'"stream":"Step 1/2 : FROM nginx:v1.0"\\r\\n'
b'"stream":"\\\\n"\\r\\n'
b'"stream":" ---\\\\u003e 5dd2557ae711\\\\n"\\r\\n'
b'"stream":"Step 2/2 : CMD [\\\\"/bin/sh\\\\"]"\\r\\n'
b'"stream":"\\\\n"\\r\\n'
b'"stream":" ---\\\\u003e Running in 6daaf3f6dd12\\\\n"\\r\\n'
b'"stream":"Removing intermediate container 6daaf3f6dd12\\\\n"\\r\\n'
b'"stream":" ---\\\\u003e af2b53a2923a\\\\n"\\r\\n'
b'"stream":"Successfully built af2b53a2923a\\\\n"\\r\\n'
b'"stream":"Successfully tagged pydocker/demo:latest\\\\n"\\r\\n'
      
——docker images #查看镜像是否创建成功

（4）对容器的操作

看了关于镜像的相关操作后，现在来看创建容器的相关操作：

#查看所有容器
import docker
client = docker.from_env()
container_list = client.containers()        # 得到所有的容器
for container_inst in container_list:
    print(container_inst['Id'Python的docker-py模块
 使用Docker Python SDK登录注册表（docker-py）
 docker-py(Docker SDK for Python)使用指南
 pythong模块turtle简单用法
 Zabbix 监控 Docker容器
 敏捷软件开发 – DIP 依赖倒置原则