rasla / docker-magic

Course materials about Docker, Docker-Compose and Kubernetes

Docker Magic

This repo contains Docker's usage examples.

I divide Docker-Magic into 3 levels:

1. Docker - using the single docker container, including topics:
   • Get docker-image;
   • Make docker-image (with Dockerfile);
   • Run docker-container;
   • Run docker-container with custom params (env-vars, port-mapping, disk-volumes, as specific user);
2. Docker-compose - using the application from several Docker-containers (on a single server), including topics:
   • Combine few docker-images into the Application
   • Configure Application (for multi envs/users)
   • Mapping files/dirs between docker-containers and Host-OS
   • Mapping TCP/UDP ports between docker-containers and Host-OS
   • Limits of the resources for docker-containers
3. Kubernetes - using the orchestrator for full control (many applications, multi-server cluster):
   • Deploy Pods (containers) in Kubernetes-cluster
   • Configure Applications by ConfigMaps, ENV-vars and Secrets
   • Expose HTTP / HTTPS routes and TCP ports for the services
   • Persistent Storage for some Pods

Requirements

• Install Docker-CE
  or deb-package docker.io (Ubuntu 16.04+, Debian 10+): sudo apt install docker.io

• (optional, but recommended) Post-installation steps for Linux:

  • Manage Docker as a non-root user
  • Configure Docker to start on boot
  • Create user-defined network: docker network create app

• To install Docker-Compose make a few steps (on Linux):

  sudo apt install python3-pip;
  sudo pip3 install docker-compose==1.27.4

YAML-linter

IMPORTANT Before commit changes - you MUST check code (& fix errors, if possible) by linter:

## (once) Install Linters
sudo pip3 install yamllint==1.28.0

## PRE-commit check
yamllint .

Production and Development advices

• Use Alpine Linux docker-images as base, if it possible (postgres:12-alpine instead of postgres:12, for example)
• Update your docker-images ASAP, when the base-image is updated
• Use .dockerignore file - prevent unnecessary files from appearing in the docker-image

Development environment

1. The developer's docker-image may exclude the source code for interpreted languages (like PHP, Python etc), if it is sufficient to build the Runtime docker-image and mount the Sources as a Volume to the docker-container.

2. Developer's docker-images may contain some useful tools, such as XDebug (for PHP), etc. that are not needed in the production docker-images.

Production environment

1. The Production docker-image MUST contain all Application files inside docker-image !
   The only exception is - connecting to the production docker-container external persistent storage for the Logs and Data (like: MySQL data-dir, shared assets-folder for Samba/NFS, etc).

2. The Production docker-image MUST NOT contain any unnecessary tools and files !

   • Less files - less docker-image size;
   • Less binary - less potential vulnerabilities