Docker container orchestration tool Compose (starting part)

In a microservice application composed of large Docker, the number of containers is very large. If you rely on traditional manual configuration for maintenance, it will be a nightmare for development and operation and maintenance. Compose emerged to solve this problem.

Compose Introduction

Compose's predecessor was Fig. After Fig was acquired by Docker, it was officially renamed Compose. Compose is backward compatible with Fig. Compose is a tool for defining and running multi-container Docker applications. It only requires a Compose configuration file and a simple command to create and run all the containers needed by the application. In the configuration file, all containers passservicesto define and usedocker-composeCommands start or stop containers and all dependencies.

Install Compose

There are many ways to install Compose, and it is recommended to use it herecurlCommand to install. Before installing, make sure that Docker is installed on your machine and can runsudo docker versionCommand to confirm whether Docker is installed. As of now, the latest release of Compose is1.11.2, the following demonstrates installing Compose on a Linux host with Docker installed.

Installation is very simple, just execute the following command:

After waiting for the installation, execute the following command,docker-composeAdd executable permissions:

chmod +x /usr/local/bin/docker-compose

enterdocker-compose --versionThe command can view the installation results.

In addition to this installation method, you can also install Compose into a Docker container through Python's pip command. For details, please refer to/compose/install/#install-as-a-container。

If you want to uninstall Compose, you can execute itsudo rm /usr/local/bin/docker-composeOrder.

Getting started with Compose

Below we demonstrate the steps of using Compose with a simple example, using Python to build a web application that uses the Flask framework and maintains a hit count in Redis (it doesn't matter even if you are not familiar with Python, you don't even need to install Python and Redis, we will get these dependencies from the container).

Create a project

First, you need a folder as the project folder:

mkdir composetest
cd composetest

Create a project folderand copy and paste the following code into the file:

from flask import Flask
from redis import Redis

app = Flask(__name__)
redis = Redis(host='redis', port=6379)

@('/')
def hello():
  count = ('hits')
  return 'Hello World! I have been seen {} times.\n'.format(count)

if __name__ == "__main__":
  (host="0.0.0.0", debug=True)

Create a project folderand copy and paste the following code into the file:

flask
redis

At this point, we have completed the work of creating new projects, coding, adding dependencies, etc.

Create a Dockerfile

Let's create oneDockerfileThe file is used to build a Docker image that contains all dependencies for running the web application, including the Python running environment.

Create a project folderDockerfileand copy and paste the following content into the file:

FROM python:3.4-alpine
ADD . /code
WORKDIR /code
RUN pip install -r 
CMD ["python", ""]

Let's explain this configuration file roughly:

1. Use python-3.4-alpine as the base image
2. Add the current directory to the image/code directory
3. Set /code as working directory
4. Install Python dependencies
5. Set default execution commands

Define services in Compose file

Create a project folderand copy and paste the following content into the file:

version: '2'
services:
 web:
  build: .
  ports:
   - "5000:5000"
  volumes:
   - .:/code
 redis:
  image: "redis:alpine"

This configuration file contains two services, namely web and redis. The web will use the Dockerfile file in the current directory to build the image, expose the 5000 port of the container to the host, and then mount the project folder to the /code directory in the container; redis will be built using the officially released image.

Build and run

Run the following command to build and run the container:

sudo docker-compose up

After the container is built and started, you can enter http://localhost:5000 in the browser to view the results. The page will print "Hello World! I have been seen 1 times." After refreshing the page, the count will accumulate to 2.

Update the application

Since the project folder is mounted in the container, we can directly modify the application of the project folder, and the modification results are immediately reflected into the container without restarting the container. Modify the return value in the hello method in the file to the following:

return 'Hello from Docker! I have been seen {} times.\n'.format(count)

After saving, refresh the browser and find that the print results have been updated.

Other commands for Compose

The above-mentioned Componse uses commands to build and start the container, which is started in the previous mode. If you want the future station to start, you can add the parameter -d, such as the following:

sudo docker-compose up -d

docker-compose psThe command can view the running container:

liuwei@liuwei-Ubuntu:~$ sudo docker-compose ps
Name           Command        State      Ports
-------------------------------------------------------------------------------------
composetest_redis_1   redis ...  Up   6379/tcp
composetest_web_1   python           Up   0.0.0.0:5000->5000/tcp

If usingsudo docker-compose up -dCommand to start the background mode, you can usedocker-compose stopThe command stops.docker-compose down --volumesThe command can stop the container and delete it.--volumnsIndicates that the redis data file directory is deleted at the same time.

More commands about Compose can be usedsudo docker-compose --helpCheck.

The above is a basic usage process of Compose. You can find that Compose willdocker runThe commands are integrated into oneIn the configuration file, it is very convenient to manage large Docker clusters, for example, you can use multipleserviceCombining into more complexservicegroup, for eachserviceSpecify differentDockerfile, and then put themlinkTogether. I hope it will be helpful to everyone's learning and I hope everyone will support me more.