Introduction to Docker

Running things in an isolated environment be it for testing, demo or simply build is such a neat thing. You don’t have to worry about other processes running on the system or ports being consumed by someone else or inconsistent libraries/packages installed. So far the only option we had was to create this heavy-weight Virtual machine that packages an entire Guest operating system with it. Due to its heavy weight nature, it was really cumbersome to distribute this 10GB+ file. There was no good way to automate the complete configuration. Also the Host you would run this on needs to be powerful enough(RAM ~ 8GB+) to be able to run this Virtual machine which often was a challenge given the kind of laptops most field folks carry. Running several such instance of VMs was impossible given the limited compute capacity of the host OS.

Come Docker Docker addresses most of these problems quite beautifully and also the developers can no more close bugs by simply saying “It works on my machine”

What is Docker?

Docker is a container virtualization technology. It is a form of light weight container that allows you to run within a host operating system. Unlike Virtual machines, you don’t package a complete Guest operating system with it. There is a thin layer however that allows you to run your application in an isolated manner over your choice of linux operating system that runs on top of the Host operating system. You can easily run several such containers on the Host Operating system. Since Docker uses underlying features(covered later) of an Linux operating system, the container needs to have Linux as their host operating system.

In a nutshell:

• Docker provides a platform that enables users to build, package, ship and run distributed applications.
• Docker users package up their applications, and any dependent libraries or files, into a Docker image.
• Docker images are portable artifacts that can be distributed across Linux environments.
• Images that have been distributed can be used to instantiate containers where applications can run in isolation from other applications running in other containers on the same host operating system.

How it is better than Virtual Machine?

• Virtual machine is a heavy weight resource that takes up significant memory and cpu resources from the Host OS.
• It comes with a copy of the entire OS.
• It runs on top of Hypervisor that takes 10-15% of resources on Host machine and even take several seconds to boot up.

Advantages as compared to Virtual Machine

• More agile and light weight compute resource.
• It launches in sub-seconds.
• The entire application get shipped in a container which is isolated from the host OS just like the VM.
• The same container can be used by several teams such as QA and DevOps.
• Since these are light-weight, you can easily spawn several such containers on the Host and simulate a clustered environment if you need.

In a typical LAMP application, you can create one or several containers for the Apache WebServer and say one for the MySQL server on same or different host and link them all together.

Usecases for Docker

• Continuous integration and continuous deployment
• Replicate production environment on developer’s own laptops
• Run Acceptance tests on the production image.
• Since you can move these containers around, you can easily given this to other teams and repeat this build QA cycle

Since container are built in seconds, you can save so much time that now you can spend your time learning something else

For installation, refer link

Registry

You can find most of the Docker images on a public repository hosted at Docker Hub. This allows you to quickly download images that are precreated or contributed by the community. e.g. ubuntu:14.04 image

The Docker client allows to search for published images from Docker Hub and download them locally that you can build containers for.

A private registry can also be created and hosted locally in order to serve images within your company.

Basic Operations

Typically you would perform the following operations:

• Search the Docker Hub or your registry for an existing image. e.g ubuntu:14.04
• Pull the specific image from the registry if not available locally
• Create Dockerfile that does additional configuration on this image like downloading more packages and some configuration.
• Create a running environment (container) from this image using the Dockerfile.
• Stop the container (optional)
• Save the container as an image
• Push the modified image back to the registry

Useful Commands

Docker commands needs to be run as root(sudo) or you can add the user to the docker group.

Search for images on dockerhub.

# Usage: docker search [image name]
$ docker search ubuntu

Look at all the images downloaded locally.

# Usage: docker images
$ docker images

Builds an image.

# Usage: docker build -t <tag> -f <file>
$ docker build -t mysql56 -f Dockerfile_mysql .

Create/Run the docker container

# Usage: docker run <options> [image name] [command to run]
$ docker run -it cfi_ubuntu:14.04 /bin/bash
$ docker run my_img echo "hello"

List running and non running containers

$ docker ps -l 

Run a named container instead of having long IDs

# Usage:  docker run -name [name] [image name] [command]
$ docker run -name my_cont_1 my_img echo "hello"

Run existing container

# Usage: docker start [container ID]
$ docker start c629b7d70666

Attach to running container

# Usage: docker exec [container ID]
$ docker exec -it c629b7d70666 bash

Stop running container

# Usage: docker stop [container ID]
$ docker stop c629b7d70666

Delete Container

# Usage: docker rm [container ID]
$ docker rm c629b7d70666

Delete Image

# Usage: docker rmi [image ID]
$ docker rmi c629b7d70666

Saving (committing) a container: This command turns your container to an image.

# Usage: docker commit [container ID] [image name]:[tag/version - default latest]
$ docker commit 8dbd9e392a96 my_new_img:v2

Underlying Technology

Docker utilizes several linux features under the hood. These are:

Namespaces

This provides us isolation from other processes running on the host which is what Docker calls as the container. Several namespaces are created for the container so that it can provide this isolated workspace.

Control Groups

This allows us to create containers with specific CPU and memory resources. This ensures that the containers behave as good citizens on the host.

Union file systems

This filesystem allows docker to create layers thereby making them fast and lightweight.

Working with Windows and Mac

The Docker daemon uses linux-specific kernel feature so you need to use Docker from a Linux machine. If you are using Windows or Mac, you can use Boot2Docker. Boot2Docker is composed of

• VirtualBox Virtual Machine (VM)
• Docker and
• Boot2Docker management tool (Lightweight Linux VM)

New Tools with Docker

Compose

There used to be a tool called fig which allowed you to manage Docker containers using a single file. Its now replaced with inbuilt tool called Compose.

Compose allows you to specify your entire application in a single yml file which could be composed of multiple containers. You can expose the correct port and create links between the containers in this file. Now you can start/stop your application using docker-compose commands instead of worrying about individual containers.

Example of docker-compose.yml

db:
  image: mysql
  expose:
    - "3306"
  volumes_from:
    - DBDATA

web:
  build: .
  working_dir: /app
  command: python manage.py runserver 0.0.0.0:8000
  volumes:
    - .:/app
  ports:
    - "8000:8000"
  links:
    - db

Commands:

Start your application # This will build the containers and run them $ docker-compose up

Stop application

$ docker-compose down

Check status of your application (containers) $ docker-compose ps

Swarm

Currently, if you run your containers across multiple hosts, most create their own customized shell script to manage cluster of Docker nodes. Instead now you could use Swarm. It is a native clustering tool for Docker. It allows your to manage your Docker nodes as a single virtual host.

It supports the following in a nutshell:

• Orchestration tool to handle several Docker nodes
• Exposes several Docker Engines as a single virtual Engine
• You can still use the standard Docker API

Swarm 0.2.0

• Supports 85% of Docker API as REST
• Management of resources such as CPU, Memory and Networking
• Multiple discovery backends
• TLS Support

Sample Docker file to build a MySQL + Django Image

FROM ubuntu:14.04

MAINTAINER John Doe

RUN apt-get update

RUN apt-get install -y build-essential git wget supervisor

# Pre-requisites for a python/pip environment
RUN apt-get install -y python python-dev python-setuptools python-pip libmysqlclient-dev
RUN easy_install pip

# Copy our runtime install components file
COPY requirements.txt /tmp

# Install all runtime packages
RUN cd /tmp; pip install -r requirements.txt

Till then…