Ken Cochrane's Docker Boston Docker 101 Talk

Ken Cochrane's Docker Boston Docker 101 Talk. It goes over the basics of Docker, and what you need to get started.

Transcript

1. Docker Boston Docker 101 Ken Cochrane 09/24/2013 1
2. About me • Ken Cochrane (@KenCochrane) • Engineer at dotCloud (Corporate sponsor of Docker) • Work on Docker and Docker related projects (index, registry, docs, etc) 2
3. Docker Swag • I have a limited amount of Docker swag to give away. • T-shirts • Stickers • Send a tweet with #DockerBoston and I’ll randomly pick winners for the T-shirts 3
4. After Bar • Lucky’s Lounge • 355 Congress Street, Boston MA 02110 • 9:00pm - ? 4
5. Outline • Intro to Docker • Common use cases • Demo • Questions 5
6. Quick survey • How many people have heard of Docker before today? • How many people have tried Docker? • How many people are using Docker on a project today? 6
7. Where did Docker come from? • Docker is a rewrite of similar code that currently powers the dotCloud PaaS • Original version written in Python, new version written in Go. • Still a very young project, but mature for it’s age. 7
8. Docker Timeline • January 2013 Docker started as internal project inside of dotCloud • March 21, 2013 Solomon gives Docker lighting talk at PyCon US • March 27, 2013 Docker released to Public • June 2013 Docker adds OpenStack compatibility • August 2013 Docker 0.6 released 8
9. In the first 6 months • 6000+ GitHub stars • 125+ contributors • 50,000+ docker index pulls • 100’s of projects built on top of Docker • UI’s, mini-PaaS, remote desktop,etc • 1000’s of Dockerized applications 9
10. What is Docker? • Docker is an open-source engine that automates the deployment of any application as a lightweight, portable, self-sufficient container that will run virtually anywhere. 10
11. How does Docker work? • LinuX Containers (LXC) • Control Groups & Namespaces • AUFS 11
12. LinuX Containers (LXC) • lets you run a Linux system within another Linux system • A container is a group of processes on a Linux box, put together in an isolated environment • Inside the box, looks like a VM. • Outside of the box, it looks like normal processes. • chroot on steroids 12
13. Why Containers? • Speed: Boots in seconds. • Footprint: 100-1000 containers on one machine. Small disk requirements 13
14. Containers vs. VMs 14
15. Why are Containers lightweight? 15
16. Installing Docker 16
17. Docker Requirements • Linux Kernel 3.8 or above • AUFS • LXC • 64Bit 17
18. Installations • Server or Desktop (Vagrant) • Docker APT repo (get.docker.io) • Source code compile • Binary download • http://docs.docker.io/en/latest/installation/ 18
19. Vagrant install • git clone https://github.com/dotcloud/docker.git • vagrant up • vagrant ssh • sudo docker 19
20. Binary install • wget http://get.docker.io/builds/Linux/x86_64/docker-latest.tgz • tar -xf docker-latest.tgz • sudo ./docker -d & • Still need to install system startup script 20
21. Docker on Digital Ocean in 5 easy steps # Create Ubuntu 13.04 64 Bit Droplet, Then $ sudo apt-get update $ sudo apt-get install linux-image-extra-`uname -r` $ sudo sh -c "curl http://get.docker.io/gpg | apt-key add -" $ sudo sh -c "echo deb https://get.docker.io/ubuntu docker main > /etc/apt/ sources.list.d/docker.list" $ sudo apt-get update && sudo apt-get install lxc-docker 21
22. Docker on Digital Ocean in 1 step When creating the Droplet. Select the Docker image under the application 22
23. Use Cases 23
24. Docker Use Cases • Local Dev environment • Deployment • Unit testing • parallelize tests • one db per test • system tests 24
25. Testing 25
26. Unit testing • Use containers to isolate tests into their own environment. • No more worrying about tests not cleaning up after themselves. • Parallelize the tests across multiple machines 26
27. System Testing • Easily create all the different system configurations to test against • No need to worry about breaking or rebuilding a test server • Test Fabric scripts • http://agiliq.com/blog/2013/06/self-testing-fabfile-using-docker/ 27
28. Continuous Integration • Run unit tests after each source commit • StriderCD.com - Open source CI server • Uses Docker Containers to run CI tests before deployment. • TravisCI is also playing with Docker 28
29. Deployment 29
30. Docker Deployment • dokku • Flynn.io • deis.io • chef, puppet, salt, ansible, etc 30
31. Dokku • Open source PaaS • Docker powered mini-heroku • Less than 100 lines of Bash • Heroku buildpacks • Git push deployment • https://github.com/progrium/dokku 31
32. Flynn.io • Open source PaaS written in Go • Uses Docker to manage containers • One of the founders is author of Dokku • Still in development 32
33. deis.io • Open source written in Python • Git push deployments • Docker images, chef recipes or Heroku buildpacks • Supports scaling and application formation 33
34. Chef • chef-docker • docker-cookbook 34
35. Puppet garethr/docker 35 35
36. Salt stack • ipmb/salt-minion • kstaken/salt-test-runner 36
37. Ansible • cove/docker-ansible • bryfry/ansible-docker 37
38. Local Development 38
39. Local Dev Setup • VM’s are heavy, containers not so much • Run 100’s of containers on laptop vs a handful of VMs • Easier to duplicate production environment if you have a complex setup. • http://blog.scoutapp.com/articles/2013/08/28/docker-git-for-deployment 39
40. Cool Projects Using Docker 40
41. Cool projects • npmt.abru.pt: Node.js Module (NPM) testing • ptone/jiffylab: Zero configuration Python/Unix web based teaching environment • kitchen-Docker: Docker driver for Ruby’s test kitchen • MemcachedAsaService.com • Try RethinkDB, openstack-docker • Many many more. 41
42. npmt.abru.pt • Autonomously Testing All NPM modules • One container per module is created then destroyed when test is finished. • 39496 modules verified 42
43. kitchen-docker • A Ruby framework for running integration tests in an isolated environment • Uses Docker to run tests in containers • https://github.com/portertech/kitchen-docker 43
44. JiffyLab • Created by Preston Holmes • Provides an entirely web based environment for instruction. • Python and UNIX shell env running in it’s own Docker container • https://github.com/ptone/jiffylab 44
45. Memcached SaaS • Built as a class project • Memcached SaaS built on top of Docker • Built with Ruby on Rails • https://github.com/jbarbier/SaaS_Memcached/ 45
46. Try RethinkDB • SaaS that let you try out RethinkDB • One DB per container • Containers killed within 24 hours • 1000’s of containers on one host 46
47. openstack-docker • Deploy lightweight linux containers on Openstack • https://github.com/dotcloud/openstack- docker 47
48. Using Docker 48
49. Docker terms • Container: Linux container • Image: a snapshot of a container that when run creates a new container. • Index: Public docker image directory • Dockerfile: An automated script used to create an Image • Push/pull : Commands used to get images to and from the index • Run: Start a Docker image to create a running Container 49
50. Common commands • ps : lists the containers on the system • images : lists the images on the system • run : runs commands against an image to create a container • stop : stops a running command • build : builds a Dockerfile • inspect: shows you information about a container • pull: pulls down a new image from the docker index • logs: shows the logs for a given container 50
51. Docker Basics 51
52. Changes and Updates 52
53. Docker Run • Run command and exit • $ docker run ubuntu echo “hello world” • Start interactive shell in container • $ docker run -i -t ubuntu bash • Run command in background • $ docker run -d ubuntu /bin/sh -c “while true; do echo hello world; date; sleep 1; done” 53
54. Dockerfiles 54
55. Dockerfiles • Simple scripting language • Automate the creation of docker images • Built in caching • Add them to any project repo to Dockerize the project. • Online tutorial • http://www.docker.io/learn/dockerfile/ 55
56. Building Images by hand # start a ubuntu 12.10 container $ docker run -i -t ubuntu:12.10 bash # update the apt repo [a2bc13] $ apt-get update # install curl [a2bc13] $ apt-get install curl # exit out of container [a2bc13] $ exit # save changes to container as an image $ docker commit -m "comment" a2bc13 username/image 56
57. Build image using Dockerfile # curl image # VERSION 0.1 FROM ubuntu:12.10 MAINTAINER yourName name@example.com RUN apt-get -qq update RUN apt-get install -y curl 57
58. Docker build • docker build . • docker build - < Dockerfile • docker build github.com/creak/docker-firefox • use “-t” to tag built image • docker build -t myname/myimage . 58
59. Redis Dockerfile example # Redis # VERSION 0.1 FROM johncosta/redis MAINTAINER yourName name@example.com EXPOSE 6379 RUN mkdir /redis ENTRYPOINT ["/usr/bin/redis-server"] CMD ["--dir", "/redis"] 59
60. Docker Index 60
61. Docker index • Similar to PyPI but for Docker images • Public directory to store and download reusable docker images • Docker image meta data • Account required to publish images (free) • Written in Django • https://index.docker.io 61
62. Docker registry • Open source Python Flask app • https://github.com/dotcloud/docker-registry • Manages the storage of the images • Install private registry for private images 62
63. Docker API 63
64. Docker API • REST based API used to control the Docker daemon • live events api feed • some websocket support • The Docker CLI uses the same API • Clients available for most languages 64
65. Docker clients • python: docker-py • ruby: docker-ruby, docker-client, docker- api • javascript: docker-js, dockerui • Java: docker-java 65
66. Docker-py import docker # create client docker_client = docker.Client(base_url='unix://var/run/docker.sock', version="1.4") # create container container = docker_client.create_container('ubuntu', None, detach=True) container_id = container.get('Id') # start Container docker_client.start(container_id) # get Ip address meta = docker_client.inspect_container(container_id) ip = meta.get('NetworkSettings').get("IPAddress") # stop container docker_client.stop(container_id) 66
67. Docker UI’s • Shipyard: Django • https://github.com/ehazlett/shipyard • Docker-UI: Angular.js • https://github.com/crosbymichael/dockerui • Dockland: Ruby • https://github.com/dynport/dockland 67
68. Demo 68
69. Coming Soon 69
70. Docker 0.7 • AUFS will be replaced with device- mapper • Red Hat Support with Fedora • Expected at the end of September 70
71. Docker 1.0 • Pluggable architecture • LXC, Solaris Zones, FreeBSD Zones, etc • AUFS, BTRFS, etc • Better OS support (RedHat, etc) 71
72. Docker 1.0 design 72
73. Want to Learn more? • Website: http://www.docker.io • Documentation: http://docs.docker.io • Github: https://github.com/dotcloud/docker • IRC: freenode #docker • Twitter: follow @docker • Google group: groups.google.com/forum/#!forum/docker- user • Meetups: Boston, New York, London, Paris, San Francisco, and more coming soon. Go to website for details. 73
74. Go Boston Meetup http://bit.ly/goboston Next Meeting Oct 1st 6-8pm @ Akamai 74
75. Questions? 75
76. Thank you! @KenCochrane 76