The challenge of application distributionIntroduction to Docker

Credits & references

Intros slides borrowed from • dotCloud’s « Docker introduction » slideshow• AWS re:Invent 2014 session APP303, « Lighting Fast Deploys with

Docker Containers and AWS »

Agenda

• The application distribution challenge• The current solutions• Introduction to Docker, Containers, and the Matrix from Hell• Why people care: Separation of Concerns• Technical Discussion• Ecosystem, momentum• How to build Docker images• How to make containers talk to each other, how to handle data

persistence• Demo 1: isolation• Demo 2: real case - installing Go Math!, tail –f containers, unit

tests

The challenge

The challenge: apps have fundamentally changed

~2000 2014

Long lived Development is iterative and constant

Monolithic and built on a single stack

Built from loosely coupled components

Deployed to a single server Deployed to a multitude of servers

Everybody already heard/said it:

Static website

Web frontend

User DB

Queue Analytics DB

Background workers

API endpoint

nginx 1.5 + modsecurity + openssl + bootstrap 2

postgresql + pgv8 + v8

hadoop + hive + thrift + OpenJDK

Ruby + Rails + sass + Unicorn

Redis + redis-sentinel

Python 3.0 + celery + pyredis + libcurl + ffmpeg + libopencv + nodejs + phantomjs

Python 2.7 + Flask + pyredis + celery + psycopg + postgresql-client

Development VM

QA server

Public Cloud

Disaster recovery

Contributor’s laptop

Production Servers

The challenge: apps have fundamentally changedM

ult

iplicit

y o

f Sta

cks

Mult

iplicit

y o

f hard

ware

envir

onm

ents

Production Cluster

Customer Data Center

Do s

erv

ices a

nd

ap

ps in

tera

ct

ap

pro

pria

tely

?

Can I m

igra

te

sm

ooth

ly a

nd

q

uic

kly

?

The Matrix From Hell

Static website

Web frontend

Background workers

User DB

Analytics DB

Queue

Dev VM QA ServerSingle Prod

ServerOnsite Cluster

Public CloudContributor’s

laptopCustomer Servers

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

The current solutions

Current solutions

• ‘Configuration’ tools like Ansible, Chef, Puppet, …

• VM tools : Vagrant

• Container tools: Google’s LMCTFY

Mult

iplicit

y o

f G

oods

Mult

ipilic

ity

of

meth

ods f

or

transp

ort

ing/sto

ring

Do I w

orry

about

how

goods

inte

ract (e

.g.

coffe

e b

eans n

ext

to s

pic

es)

Can I tra

nsport

quic

kly

and

sm

ooth

ly(e

.g. fro

m b

oat to

tra

in to

truck)

Cargo Transport Pre-1960

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

? ? ? ? ? ? ?

Also a matrix from hell

Mult

iplicit

y o

f G

oods

Mult

iplicit

y o

f m

eth

ods f

or

transport

ing/sto

rin

g

Do I w

orry

about

how

goods

inte

ract (e

.g.

coffe

e b

eans n

ext

to s

pic

es)

Can I tra

nsport

quic

kly

and

sm

ooth

ly(e

.g. fro

m b

oat to

tra

in to

truck)

Solution: Intermodal Shipping Container

…in between, can be loaded and unloaded, stacked, transported efficiently over long distances, and transferred from one mode of transport to another

A standard container that is loaded with virtually any goods, and stays sealed until it reaches final delivery.

Containers applied to software

Static website Web frontend User DB Queue Analytics DB

Dev VMQA server Public Cloud

Contributor’s laptop

Docker is a shipping container system for code M

ult

iplicit

y o

f Sta

cks

Mult

iplicit

y o

f hard

ware

envir

onm

ents

Production Cluster

Customer Data Center

Do s

erv

ices a

nd

apps in

tera

ct

appro

pria

tely

?

Can I m

igra

te

sm

ooth

ly a

nd

quic

kly

…that can be manipulated using standard operations and run consistently on virtually any hardware platform

An engine that enables any payload to be encapsulated as a lightweight, portable, self-sufficient container…

Static website

Web frontend

Background workers

User DB

Analytics DB

Queue

Dev VM QA ServerSingle Prod

Server

Onsite

ClusterPublic Cloud

Contributor’s

laptop

Customer

Servers

Docker eliminates the matrix from Hell

Good for developers

• Build once…(finally) run anywhere*• A clean, safe, hygienic and portable runtime environment for your app.• No worries about missing dependencies, packages and other pain points during

subsequent deployments.• Run each app in its own isolated container, so you can run various versions of

libraries and other dependencies for each app without worrying• Automate testing, integration, packaging…anything you can script • Reduce/eliminate concerns about compatibility on different platforms, either your

own or your customers. • Cheap, zero-penalty containers to deploy services? A VM without the overhead of a

VM? Instant replay and reset of image snapshots? That’s the power of Docker

* Anywhere = any x86 server running a modern Linux kernel (3.2+ generally. 2.6.32+ for RHEL 6.5+, Fedora, & related)

Good for (Dev)Ops

• Configure once…run anything• Make the entire lifecycle more efficient, consistent, and repeatable• Increase the quality of code produced by developers. • Eliminate inconsistencies between development, test, production, and customer

environments• Support segregation of duties• Significantly improves the speed and reliability of continuous deployment and

continuous integration systems• Because the containers are so lightweight, address significant performance, costs,

deployment, and portability issues normally associated with VMs

Why it works—separation of concerns

• Dan the Developer– Worries about what’s “inside” the

container• His code• His Libraries• His Package Manager• His Apps• His Data

– All Linux servers look the same

• Oscar the Ops Guy• Worries about what’s

“outside” the container• Logging

• Remote access

• Monitoring

• Network config

• All containers start, stop, copy, attach, migrate, etc. the same way

Under the hood

• To learn more about how it works, some background:

–Linux namespaces• PID• IPTables• UTS (~hostname)• Mount• IPC (inter-process)

–CGroups–LXC Containers

AppA

Containers vs. VMs

Hypervisor (Type 2)

Host OS

Server

GuestOS

Bins/

Libs

AppA’

Guest

OS

Bins/

Libs

AppB

Guest

OS

Bins/

Libs

App A

’

Docker

Host OS

Server

Bins/Libs

App A

Bins/Libs

App B

App B

’

App B

’

App B

’VM

Container

Containers are isolated,but share OS and, whereappropriate, bins/libraries

GuestOS

GuestOS

…result is significantly faster deployment, much less overhead, easier migration, faster restart

Why are Docker containers lightweight?

Bins/

Libs

AppA

Original App(No OS to takeup space, resources,or require restart)

App

ΔBin

s/

AppA

Bins/

Libs

AppA’

Guest

OS

Bins/

Libs

Modified App

Copy on write capabilities allow us to only save the diffs between container A and container A’

VMsEvery app, every copy of anapp, and every slight modificationof the app requires a new virtual server

AppA

GuestOS

Bins/

Libs

Copy of AppNo OS. CanShare bins/libs

AppA

GuestOS

GuestOS

VMs Containers

Containers, images, …

Terminology

Docker concept OOP analogy

Dockerfile Source code

Image Class

Container Instance

What are the basics of the Docker system?

Source Code

Repository

DockerfileFor

A

Docker Engine

DockerContainer

Image Registry

Build

Docker

Host 2 OS (Linux)

Conta

iner A

Conta

iner B

Conta

iner C

Conta

iner A

Push

SearchPull

Run

Host 1 OS (Linux)

Changes and Updates

Docker Engine

DockerContainer

Image Registry

Docker Engine

Push

Update

Bins/

Libs

AppA

App

Δ

Bin

s/

Base Container

Image

Host is now running A’’

Container Mod A’’

App

ΔBin

s/

Bins/

Libs

AppA

Bin

s/

Bins/

Libs

AppA’’

Host running A wants to upgrade to A’’. Requests update. Gets only diffs

Container Mod A’

A new open source project… or the next big thing?

• Compatible with Jenkins, Travis, Chef, Puppet, Vagrant and OpenStack…

• Ecosystem: lots of tools, even by serious contributors (Google –Kubernetes, cAdvisor - NetFlix …), new IAAS provider emerge (like Tutum),

• Supported by Amazon, Google, Microsoft, Rackspace, Digital Ocean, Joyent, RedHat, …

• Momentum... See pull requests numbers and frequency to get an idea

• Used in production… even by big players (HP, Ebay, New Relic, Etsy, Yelp, Spotify, …)

Ecosystem Support

• Operating systems– Virtually any distribution with a 2.6.32+ kernel– Red Hat/Docker collaboration to make work across RHEL 6.4+, Fedora, and other members of the family (2.6.32

+)– CoreOS—Small core OS purpose built with Docker

• OpenStack– Docker integration into NOVA (& compatibility with Glance, Horizon, etc.) accepted for Havana release

• Private PaaS– OpenShift– Solum (Rackspace, OpenStack) – Other TBA

• Public PaaS– Deis, Voxoz, Cocaine (Yandex), Baidu PaaS

• Public IaaS– Native support in Rackspace, Digital Ocean,+++– AMI (or equivalent) available for AWS & other

• DevOps Tools– Integrations with Chef, Puppet, Jenkins, Travis, Salt, Ansible +++

• Orchestration tools– Mesos, Heat, ++– Shipyard & others purpose built for Docker

• Applications– 1000’s of Dockerized applications available at index.docker.io

Use Cases

• Ted Dziuba on the Use of Docker for Continuous Integration at Ebay Now– https://speakerdeck.com/teddziuba/docker-at-ebay– http://www.youtube.com/watch?feature=player_embedded&v=0Hi0W4gX--4

• Sasha Klizhentas on use of Docker at Mailgun/Rackspace

• http://www.youtube.com/watch?feature=player_embedded&v=CMC3xdAo9RI

• Sebastien Pahl on use of Docker at CloudFlare

• http://www.youtube.com/watch?feature=player_embedded&v=-Lj3jt_-3r0

• Cambridge HealthCare• http://blog.howareyou.com/post/62157486858/continuous-delivery-with-docker-

and-jenkins-part-i

• Red Hat Openshift and Docker

• https://www.openshift.com/blogs/technical-thoughts-on-openshift-and-docker

Use Cases—From the Community

Use Case Examples LinkClusters Building a MongoDB cluster using docker http://bit.ly/1acbjZf

Production Quality MongoDB Setup with Docker http://bit.ly/15CaiHbWildfly cluster using Docker on Fedora http://bit.ly/1bClX0O

Build your own PaaS OpenSource PaaS built on Docker, Chef, and Heroku Buildpacks http://deis.io

Web Based Environment for Instruction

JiffyLab – web based environment for the instruction, or lightweight use of, Python and UNIX shell

http://bit.ly/12oaj2K

Easy Application Deployment

Deploy Java Apps With Docker = Awesome http://bit.ly/11BCvvuHow to put your development environment on docker http://bit.ly/1b4XtJ3Running Drupal on Docker http://bit.ly/15MJS6B

Installing Redis on Docker http://bit.ly/16EWOKh

Create Secure Sandboxes

Docker makes creating secure sandboxes easier than ever http://bit.ly/13mZGJH

Create your own SaaS Memcached as a Service http://bit.ly/11nL8vh

Automated Application Deployment

Multi-cloud Deployment with Docker http://bit.ly/1bF3CN6

Continuous Integration and Deployment

Next Generation Continuous Integration & Deployment with dotCloud’sDocker and Strider

http://bit.ly/ZwTfoy

Testing Salt States Rapidly With Docker http://bit.ly/1eFBtcm

Lightweight Desktop Virtualization

Docker Desktop: Your Desktop Over SSH Running Inside Of A Docker Container

http://bit.ly/14RYL6x

How to build an image?

Two ways to create an image:

• You work in a container, then commit it for exploratory work

• You build a new image from a Dockerfile for automated, repeatable process

1 container ?= (1process|1task|other) ?

Best practices appear

• Do not force you to 1 process per container• Ideal case: 1 ‘task’ per container, with separation of

concerns.

• The task might require 1 single process•Or 0 (for data-only containers)•Or several (ex: nginx+php-fpm, monitoring, …)

Containers are isolated… then how to make them talk to each other?

• Private network between containers• Linking to make them easy to talk to a friend container

• Expose ports•Mount volumes (rw or ro)

• Advanced/not so trivial other solutions–Service discovery–Enter into the container’s namespace

How to handle data persistence

• Data only containers, sharing data volumes•Mounting this data volumes in the running containers

How to handle data persistence

How data volume works

How to handle data persistence

How data volume works

Demo time!

•Quick look at the Docker commands

• Isolation demo

•Multi containers demo: installing and launching Go Math! Academy in seconds.

• Running unit tests in a container

Warning

•Docker is great–It reduces dependencies, provides isolated and controlled context

•But it brings new tools, and new technologies, hence a potential complexity increase–libswarm, systemd, etcd, ambassadord, fleet, etc

Some references

• Start by the online tutorial and documentation, they are great!

• Here’s some links kept while learning… but so much available!–Eight Docker Development Patterns–Panamax and Docker and Some Best Practices–Tutum blog–The 5 Most Important Things I’ve Learned From Using Docker

–Understanding Docker Volumes–Docker's Youtube videos

More references

First steps• dotScale 2013 - Solomon Hykes - Why we built Docker (video)• Twitter University – Solomon Hykes – Introduction to Docker (video)

Going further• Docker layers cost (blog), on how to optimize your images layer size.• Twitter University - Docker at Spotify (video)• Top 10 open-source Docker developer tools (blog, incomplete but gives ideas)• Docker In-depth: Volumes (blog)

Food for thought / Stimulate your imagination• AWS re:Invent 2014 - Lightning fast deploys with Docker containers and AWS (video,

slides)• FutureStack14 - Docker and the Future of Modern Software (New Relic blog, video) • AWS re:Invent 2014 - Infrastructure as Code (video, more related to Cloudformation

than to Docker but very inspiring if you consider the two technologies)• Alternative? Rocket by CoreOS… started a debate (like Joyent answer)

Tools for dev, day-to-day

• For everyone

• On Mac– Kitematic UI to manage your images/containers

• On Windows– Try the new XShell5 beta to easily deal with all your current containers

• So many others… and new ones every week

• Fig (Fast isolated environments, by Docker) repo, site• A Better Development Environment with Docker and Fig (blog) • Lots of tools to manage multi containers apps: Kubernetes, Shipyard,

Panamax,

Thank you