Why Docker? A Practical Guide to Containers, Dockerfile & Compose

Why Docker?

Traditional web services often depend on many programs (PHP, MySQL, Redis, Node, etc.) that must be installed manually, leading to version conflicts, shared configuration file modifications, and high manual deployment effort.

Different services may require different versions of the same software (e.g., Python2 vs Python3, PHP7 vs PHP5.6).

Multiple services can modify the same configuration files, causing interference.

Manual installation on each machine wastes considerable labor.

Packaging an entire system into a virtual machine solves some issues but introduces new problems: large image size, slow startup, and lack of automation.

VM images contain the whole OS, making them very large.

Services must wait for the OS to boot before they can start.

The packaging process cannot be fully automated.

Docker, as a process‑level virtualization technology, offers several advantages:

Small image size because the host OS is not packaged.

Fast startup with low resource consumption.

Sandbox isolation ensures environment separation between services.

Dockerfile enables automated image building.

Docker Hub provides a convenient image sharing platform.

Below is a visual comparison of VM and Docker technologies (VM packages a guest OS, Docker virtualizes directly on the host).

Docker Basics

Docker runs on Windows, Linux, macOS, AWS, and Azure (Windows requires Win10+, macOS requires El Capitan+). It follows a client/server architecture; after installation, the Docker daemon must be started before using Docker commands.

Docker has four core concepts: Dockerfile, Image, Container, and Repository. An Image is built from a Dockerfile, can be pushed to Docker Hub, and instantiated as a Container. A simple command to run an image is:

$ docker run [organization/]<image_name>[:tag]

If the organization is omitted, it defaults to library; if the tag is omitted, it defaults to latest. For example, running the hello-world image pulls it from Docker Hub if not present locally and then executes its entrypoint.

Dockerfile

Dockerfile is the core of Docker image creation. It provides two main benefits:

Text‑based image generation enables version control and automated deployment.

Each instruction creates a new layer, allowing incremental updates and layer reuse, which reduces image size.

Key Dockerfile directives include:

Use # for comments. FROM specifies the base image. RUN executes commands during image build (e.g., installing packages). COPY copies files into the image. WORKDIR sets the working directory. CMD or ENTRYPOINT defines the container’s startup command (only one can exist; ENTRYPOINT is generally preferred).

Example Dockerfile for a simple Node script:

FROM mhart/alpine-node:8.9 LABEL maintainer="lizheming <[email protected]>" WORKDIR /wechat COPY package.json /wechat/package.json RUN npm install --production --registry=https://registry.npm.taobao.org COPY index.js /wechat/index.js ENTRYPOINT ["node", "/wechat/index.js"]

Build the image with:

$ docker build lizheming/drone-wechat:latest

Docker Compose

Complex projects often require multiple services, and a single container’s ENTRYPOINT cannot handle them all. Docker Compose, a Python‑based orchestration tool, simplifies this by defining all services in a YAML file and managing them with docker‑compose up and docker‑compose stop.

Install Compose via pip: pip install docker-compose Sample docker-compose.yaml:

version: "2" services: nginx: depends_on: - php image: "nginx:latest" volumes: - "$PWD/src/docker/conf:/etc/nginx/conf.d" - "$PWD:/home/q/system/m_look_360_cn" ports: - "8082:80" container_name: "m.look.360.cn-nginx" php: image: "lizheming/php-fpm-yaf" volumes: - "$PWD:/home/q/system/m_look_360_cn" container_name: "m.look.360.cn-php"

Compose also injects service names into each container’s /etc/hosts, allowing direct name‑based access (e.g., php:9000 in an Nginx config).

Docker Related

Beyond simplifying deployment, Docker enables:

CI/CD pipelines for continuous integration and delivery.

Elastic scaling with Kubernetes or Docker Swarm, automatically adjusting resources based on load.

Learning Resources

What is Docker? https://cloud.tencent.com/developer/article/1005172

Docker from Beginner to Practice https://yeasy.gitbooks.io/docker_practice/

Docker Compose Detailed Guide https://www.jianshu.com/p/2217cfed29d7

Deep Dive into Docker https://www.kancloud.cn/infoq/docker/79768