namgivu / docker-tutorial

Docker tutorial

Prerequisite

• Docker CE
• Docker-compose

Note: Docker compose might already be included within Docker CE. So after install Docker CE, you should run docker-compose version to check if it was installed or not.

To check if installation was success, run these command

docker version
docker-compose version

Slide

https://www.dropbox.com/s/fh3feiht35l5xx6/docker.pptx?dl=0

App Description

This is a simple api with express js and mongo db

Endpoints:

• GET /: should return "Hello world".
• GET /list: should return a list of item
• GET /list/add/:name: should add a new item to list

Table of Content

• What is Docker?
• Basic concepts of Docker
• Running an application using Docker
• Containerize your application

What is Docker?

Describe the situation

• Your Company start with a simple stack:
  • Apache + MySQL + PHP
• Several months later, your boss require to have a new stack:
  • Python tensor flow
• You need to upgrade your environment
  • From Python 2 to Python 3

After a while, your environment will look like this

The Problem

in brief Multiple apps shared same OS environment and it's hard to change/install new env and/or new app

in full

• Time consuming when setup a new stack

• Environment conflict between stacks (not isolated) eg nodejs installed and broken by upgrading other stuff by not isolated it means not isolated stacks aka. stacks shared same dependencies

• Missing dependencies

• Hard to upgrade (code + environment)

• Hard to replicate to another machine

• ...

Introduction to Docker

• Docker was introduced in 2013
• For easier to understand, Docker is similar to VM but dont consume as much computing resources
• Docker allow application run in an isolated environment
• Docker provide an easy way to install and share your application

Basic concept of Docker

• Docker image: a snapshot of current machine (As concept it is similar to a GHOST Image / Backup Image of computer)
• Docker container: a running Image
• Docker registry: a centralize repository of Docker Image

Practice

• Running a Docker container

# Run a container
docker run -d                            --name mongo        mongo
           # run as background daemon    # container name    # image name at hub.docker.com/_/mongo

# List running containers
docker ps
sample_output='
CONTAINER ID        IMAGE    COMMAND                  CREATED             STATUS              PORTS                                                                                         NAMES
fb93ea67f4f1        mongo    "docker-entrypoint.s…"   4 minutes ago       Up 4 minutes        27017/tcp
'

# View logs of a container
docker logs mongo

# Stop a container
docker stop mongo

# Remove a container
docker rm mongo

# View cpu/memory consumed
docker stats  # similar to htop

Docker-compose

• A Configuration-way for running docker containers

version: "3.7"

services:
  db:
    image: mongo
    container_name: "mongo"

• Running:

# Running a single service
docker-compose up -d mongo

# Running all services
docker-compose up -d

Fundamental of Docker environment

• Two entity:
  • Docker environment: environment of Docker application
  • Host OS environment: environment of the physical server running Docker
• Docker application run in a ISOLATED ENVIRONMENT with server host
  • Isolated network
  • Isolated file system
• Docker environment is DESTROYED whenever application stopped

• Docker provide 2 method for communicate between Docker environment and host OS
  • Port Mapping: For network communication
  • Volume Mount: For file system communication

Advice

• Should limit 1 service/app per container, avoid adding multiple services into 1 i.e. Docker auto-restart a crashed container --> 1app-1container will help keeping all app alive

• Docker have its own mechanism for process monitoring and restart, so better avoid adding to the container monitor-things like supervisord, pm2, ...

Practice

docker run -d --name mongo -p 27017:27017 -v ./data/mongo:/data/db mongo

Containerize your application

• Creating Dockerfile and .dockerignore

• Running docker build -t {image_name}:{image tag} .

• Docker image is stored as multiple layer

  • Docker will reuse old layer if command or content does not change e.g. FROM node:alpine will be stored as a layer and reused whenever image node:alpine is required and this FROM command marked with ---> Using cache

  • layer for COPY command is considered new layer if file content is new

  • layer for RUN command is considered new layer if command call's whole text is new

• Reuse layer for faster docker build process

  • carefully order your dockerfile commands so that maximize the number of reused layers will give you the fastest image build time

Summary