Master Docker: From Basics to Building Your Own Images

Quick Start

Docker began as an internal side project at dotCloud.

It is written in Go.

The goal is a lightweight OS‑level virtualization solution.

Docker builds on Linux container technologies such as LXC.

Containers start in seconds, far faster than traditional VMs.

High resource efficiency allows thousands of containers on a single host.

The following images compare Docker with traditional virtualization: containers share the host OS kernel, while VMs virtualize the hardware layer.

Containers consume minimal extra system resources, delivering high performance with low overhead. Running ten applications requires ten VMs in a traditional setup, but only ten isolated containers with Docker.

Main advantages:

Faster delivery and deployment – containers are the smallest unit.

More efficient kernel‑level virtualization.

Easier migration and scaling.

Simpler management.

Installation

The official website provides installation guides for Mac, Linux, and Windows. Follow the official documentation.

Kitematic (Visual Docker Container Management on Mac & Windows) offers a graphical interface for beginners.

Daemon

When running the Docker daemon you can change the binding interface with the -H flag, e.g.: sudo /usr/bin/docker -d -H tcp://0.0.0.0:2375 To avoid typing the long command each time, set the environment variable:

export DOCKER_HOST="tcp://0.0.0.0:2375"

Web UI

Common web management tools include Shipyard and Portainer.

Basic Concepts

Image

A read‑only template used to create containers.

Based on a Union file‑system layered structure.

Can be built, updated, or pulled from public repositories.

Container

A running instance created from an image, with a writable top layer.

Can be started, stopped, and deleted; each container is isolated.

Essentially a lightweight Linux environment with its own process, user, and network spaces.

Registry

A centralized storage for images, public or private.

The largest public registry is Docker Hub.

Users push their own images to a registry and pull them on other machines.

Conceptually similar to Git repositories.

Docker follows a client/server architecture; the client sends requests to the Docker daemon.

Typical Use Cases

Standardize, optimize, and accelerate local development and build pipelines.

Ensure consistent runtime results across environments.

Create isolated environments for testing.

Isolation Features

File‑system isolation – each container has its own root FS.

Process isolation – containers run in separate process spaces.

Network isolation – separate virtual interfaces and IP addresses.

Resource isolation via cgroups – CPU, memory, etc., are allocated per container.

Common Commands

Check Docker status: sudo docker info List running containers: docker ps (add -a for all, -l for the latest)

Simple Example

Run an Ubuntu container and print a message:

docker run ubuntu:14.04 /bin/echo 'Hello wdx!'

Start an interactive Bash session: docker run -t -i ubuntu:14.04 /bin/bash Options: -t allocates a pseudo‑TTY. -i keeps STDIN open.

Inside the container you can run ps or top to see only the processes you started.

Image Management

Pull an image: docker pull ubuntu:12.04 Pull from a specific registry: docker pull dl.dockerpool.com:5000/ubuntu:12.04 List local images:

docker images

Running an Image

Search Docker Hub for the whalesay image and run it:

docker run docker/whalesay cowsay boo

Building Your Own Image

Create a directory and enter it: mkdir wdxtub; cd wdxtub Create a Dockerfile (e.g., with gedit Dockerfile).

Start from a base image: FROM docker/whalesay:latest Add commands, e.g., RUN apt-get -y update && apt-get install -y fortunes.

Specify the default command: CMD /usr/games/fortune -a | cowsay.

Build the image: sudo docker build -t wdx-whale . Verify with docker images.

Run the custom image:

docker run wdx-whale

Pushing to Docker Hub

Create a repository (e.g., wdxtub/demo) on Docker Hub, tag the image, log in, and push:

docker tag 26ac9649d7da wdxtub/wdx-whale:latest

docker login --username=wdxtub [email protected]

docker push wdxtub/wdx-whale

Exporting and Importing Images

Export an image to a tar file: docker save -o wdx-local-whale.tar wdxtub/wdx-whale Import it later:

docker load --input wdx-local-whale.tar

Running Containers in Background

Use -d to run a container detached:

docker run -d ubuntu:14.04 /bin/sh -c "while true; do echo hello world; sleep 1; done"

View logs with docker logs <container_id>. The container’s lifetime depends on the command executed, not on the -d flag.

Stopping and Restarting

Stop a container: docker stop <container_id> Restart a container:

docker restart <container_id>

Import/Export and Deletion

Export a container’s filesystem: docker export <container_id> Import a snapshot as an image: cat ubuntu.tar | sudo docker import - test/ubuntu:v1.0 Delete containers (add -f to force removal of running containers):

docker rm $(docker ps -a -q)

Repositories vs. Registries

A repository stores images; a registry is the server that hosts one or more repositories. For example, in dl.dockerpool.com/ubuntu, dl.dockerpool.com is the registry and ubuntu is the repository.