Giter Club home page Giter Club logo

docket's Introduction

Docket

DOCKEr + torrenT = Docket

Docket is a custom docker registry that allows for deployments through bittorrent. It allows for lightning fast Docker image deploys across large number of machines.

This was designed and built in 48 hours as part of the Gopher Gala Golang 48 hour hackathon. Hence kindly forgive me for the hackish code, and lack of tests.

[On hind sight - container rush or docker flood would have been a better name.]

Screencast

Problem Statement

As containers get more mainstream, server deployments are soon going to be done through a containerization solution like docker [or rocket or lxd].

Large scale deploys are going to choke your docker registry. Imagine pulling/deploying a 800mb base image (for example the official perl image) across 200 machines in one go. That's 800*200 = 1.6TB of data that's going to be deployed and it'll definitely choke your private docker registry (and will take a while pulling in from the public image)

Solution

The solution is BitTorrent. Any technology good for distributing music and movies, is usually good for doing server deploys. Instead of downloading from one point source, we can now distribute the load across the machines downloading themselves. All of this transparently and automatically handled by the docket client.

Silly Totally Non-scientific Benchmark

Pulling a 800 mb image across 5 machines simultaneously via Docket resulted in a 90% speedup.

Features

  • Written in Golang :)
  • Very easy to use
  • Works along side your private docker registry

Concepts

Docket constitutes of 3 components:

  1. Docket Registry

A REST service that acts as a registry. It receives docker image tarballs from the client, stores metadata into a database, creates torrents out of them and seeds them.

  1. Docket Client

The client itnerface in which the end user will be interacting. Can view available images in the registry, push an image to the registry and pull an image (which triggers a bittorrent deploy) from the registry.

  1. Bittorrent Tracker

Docket allows you to BYOT (bring your own tracker), but I suggest installing opentracker

Docket uses ctorrent for the actual downloading of torrents.

Installation

Step 0: You need to already have Docker installed

Step 1: Install ctorrent and opentracker

$ sudo apt-get update
$ sudo apt-get install zlib1g-dev make g++ ctorrent
$ git clone https://github.com/gebi/libowfat
$ make
$ git clone git://erdgeist.org/opentracker
$ cd opentracker
$ make

Step 2: Download the Docket binaries

$ wget http://storage.googleapis.com/docket/docket.zip
$ unzip docket.zip

Step 3: That's all

Getting Started

Step 1: Fire up the Tracker

$ cd opentracker
$ ./opentracker 8940

Step 2: Fire up the Registry

$ cd registry
# Note you need to put an ip address in which other machines can contact the tracker
# You cannot put in localhost or 127.0.0.1
$ sudo ./registry --tracker "10.240.101.85:8940"

Usage

Push an image to the Registry

$ cd client
#Note: You will have to to explicitly mention the tag ":latest"
$ sudo ./client -h "http://10.240.101.85" push netvarun/test:latest
Found image:  netvarun/test:latest
ID:  353b94eb357ddb343ebe054ccc80b49bb6d0828522e9f2eff313406363449d17
RepoTags:  [netvarun/test:latest]
Created:  1422145581
Size:  0
VirtualSize:  188305056
ParentId:  d7d8be71d422a83c97849c4a8e124fcbe42170d5ce508f339ce52be9954dc3b4
Exporting image to tarball...
Successively exported tarball...
key =  image  val =  netvarun/test:latest
key =  id  val =  353b94eb357ddb343ebe054ccc80b49bb6d0828522e9f2eff313406363449d17
key =  created  val =  1422145581
Successfully uploaded image:  netvarun/test:latest  to the Docket registry.

List all images available in the Registry

$ cd client
$ sudo ./client -h "http://10.240.101.85" images
netvarun/test:latest
perl:latest

Do a deploy

$ cd client
$ sudo ./client -h "http://10.240.101.85" pull perl:latest
Downloading the torrent file for image:  perl:latest
Downloading from the torrent file...
META INFO
Announce: http://10.240.101.85:8940/announce
Piece length: 524288
Created with: docket-registry
FILES INFO
<1> 14f61693dd2db6380755a662d6e4e3583b5214fad9032bd983ce6c70df2144bc_perl_latest.tar [838467072]
Total: 799 MB
Found bit field file; verifying previous state.
Listening on 0.0.0.0:2706
Press 'h' or '?' for help (display/control client options).
Checking completed.
- 0/0/1 [0/1600/0] 0MB,0MB | 0,0K/s | 0,0K E:0,0 Connecting
End of input reached.
Input channel is now off
\ 1/0/2 [1572/1600/1600] 785MB,0MB | 36422,0K/s | 38896,0K E:0,1
Download complete.
Total time used: 0 minutes.
Seed for others 0 hours
| 0/0/2 [1600/1600/1600] 799MB,0MB | 0,0K/s | 14320,0K E:0,1 Connecting

Tarball path =  /tmp/docket/14f61693dd2db6380755a662d6e4e3583b5214fad9032bd983ce6c70df2144bc_perl_latest.tar
Exporting image to tarball...

Reference (Docket Registry)

./registry --help

usage: registry [<flags>]

Docket Registry

Flags:
  --help             Show help.
  -t, --tracker="10.240.101.85:8940"  
                     Set host and port of bittorrent tracker. Example: -host 10.240.101.85:8940 Note: This cannot be set to localhost, since this is the tracker in which all the torrents will be created with. They have to be some accessible ip
                     address from outside
  -p, --port="8000"  Set port of docket registry.
  -l, --location="/var/local/docket"  
                     Set location to save torrents and docker images.

Reference (Docket Client)

./client 

usage: client [<flags>] <command> [<flags>] [<args> ...]

Docket Client

Flags:
  --help             Show help.
  -h, --host="http://127.0.0.1"  
                     Set host of docket registry.
  -p, --port="8000"  Set port of docket registry.
  -l, --location="/tmp/docket"  
                     Set location to store torrents and tarballs.

Commands:
  help [<command>]
    Show help for a command.

  push <push>
    Push to the docket registry.

  pull <pull>
    pull to the docket registry.

  images [<flags>]
    display images in the docket registry.

Author

Sivamani Varun ([email protected])

License

MIT License

docket's People

Contributors

netvarun avatar

Watchers

 avatar  avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.