大多数情况下, 完整的应用由多个服务组合而成, docker解决了单容器(服务)的问题, 但无法同时协调多个容器, docker-compose正是解决这一问题
Unfortunately, when you are a developer you rarely create a stand-alone program (a program that does not require any other services to run, such as a database).
However, how do you know if you need Docker-Compose? It’s very simple — if your application requires several services to run, you need this tool. For example, if you create a website that needs to connect to your database to authenticate users (here 2 services, website and database).
Docker-compose offers you the possibility to launch all these services in a single command.
. ├── client/ ├── docker-compose.yml └── server/
Dockerfile + 业务代码
. ├── Dockerfile ├── index.html └── server.py
server.py内容如下, 创建一个简单的web服务, 监听端口1234
#!/usr/bin/env python3 # Import of python system libraries. # These libraries will be used to create the web server. # You don't have to install anything special, these libraries are installed with Python. import http.server import socketserver # This variable is going to handle the requests of our client on the server. handler = http.server.SimpleHTTPRequestHandler # Here we define that we want to start the server on port 1234. # Try to remember this information it will be very useful to us later with docker-compose. with socketserver.TCPServer(("", 1234), handler) as httpd: # This instruction will keep the server running, waiting for requests from the client. httpd.serve_forever()
Docker-Compose is magic!
FROM python:latest ADD server.py /server/ ADD index.html /server/ # I would like to introduce something new, the 'WORKDIR' command. # This command changes the base directory of your image. # Here we define '/server/' as base directory (where all commands will be executed). WORKDIR /server/
. ├── client.py └── Dockerfile
#!/usr/bin/env python3 # Import of python system library. # This library is used to download the 'index.html' from server. # You don't have to install anything special, this library is installed with Python. import urllib.request # This variable contain the request on 'http://localhost:1234/'. # You must wondering what is 'http://localhost:1234'? # localhost: This means that the server is local. # 1234: Remember we define 1234 as the server port. fp = urllib.request.urlopen("http://localhost:1234/") # 'encodedContent' correspond to the server response encoded ('index.html'). # 'decodedContent' correspond to the server response decoded (what we want to display). encodedContent = fp.read() decodedContent = encodedContent.decode("utf8") # Display the server file: 'index.html'. print(decodedContent) # Close the server connection. fp.close()
FROM python:latest ADD client.py /client/ WORKDIR /client/
# A docker-compose must always start by the version tag. # We use "3" because it's the last version at this time. version: "3" # You should know that docker-composes works with services. # 1 service = 1 container. # For example, a service maybe, a server, a client, a database... # We use the keyword 'services' to start to create services. services: # As we said at the beginning, we want to create: a server and a client. # That is two services. # First service (container): the server. # Here you are free to choose the keyword. # It will allow you to define what the service corresponds to. # We use the keyword 'server' for the server. server: # The keyword "build" will allow you to define # the path to the Dockerfile to use to create the image # that will allow you to execute the service. # Here 'server/' corresponds to the path to the server folder # that contains the Dockerfile to use. build: server/ # The command to execute once the image is created. # The following command will execute "python ./server.py". command: python ./server.py # Remember that we defined in'server/server.py' 1234 as port. # If we want to access the server from our computer (outside the container), # we must share the content port with our computer's port. # To do this, the keyword 'ports' will help us. # Its syntax is as follows: [port we want on our machine]:[port we want to retrieve in the container] # In our case, we want to use port 1234 on our machine and # retrieve port 1234 from the container (because it is on this port that # we broadcast the server). ports: - 1234:1234 # Second service (container): the client. # We use the keyword 'client' for the server. client: # Here 'client/ corresponds to the path to the client folder # that contains the Dockerfile to use. build: client/ # The command to execute once the image is created. # The following command will execute "python ./client.py". command: python ./client.py # The keyword 'network_mode' is used to define the network type. # Here we define that the container can access to the 'localhost' of the computer. network_mode: host # The keyword'depends_on' allows you to define whether the service # should wait until other services are ready before launching. # Here, we want the 'client' service to wait until the 'server' service is ready. depends_on: - server
$ docker-compose build $ docker-compose up