Backup is a RubyGem, written for Linux and Mac OSX, that allows you to easily perform backup operations on both your remote and local environments. It provides you with an elegant DSL in Ruby for modeling your backups. Backup has built-in support for various databases, storage protocols/services, syncers, compressors, encryptors and notifiers which you can mix and match. It was built with modularity, extensibility and simplicity in mind.

Build Status Still Maintained


Michael van Rooijen ( @meskyanichi )

Drop me a message for any questions, suggestions, requests, bugs or submit them to the issue log.


To get the latest stable version

gem install backup

You can view the list of released versions over at (Backup)

Getting Started

I recommend you read this README first, and refer to the wiki pages afterwards. There's also a Getting Started wiki page.

What Backup 3 currently supports

Below you find a list of components that Backup currently supports. If you'd like support for components other than the ones listed here, feel free to request them or to fork Backup and add them yourself. Backup is modular and easy to extend.

Database Support

  • MySQL
  • PostgreSQL
  • MongoDB
  • Redis
  • Riak

Database Wiki Page

Filesystem Support

  • Files
  • Directories

Archive Wiki Page

Storage Locations and Services

  • Amazon Simple Storage Service (S3)
  • Rackspace Cloud Files (Mosso)
  • Ninefold Cloud Storage
  • Dropbox Web Service
  • Remote Servers (Available Protocols: FTP, SFTP, SCP and RSync)
  • Local Storage

Storage Wiki Page

Storage Features

  • Backup Cycling, applies to:
    • Amazon Simple Storage Service (S3)
    • Rackspace Cloud Files (Mosso)
    • Ninefold Cloud Storage
    • Dropbox Web Service
    • Remote Servers (Only Protocols: FTP, SFTP, SCP)
    • Local Storage
  • Backup Splitting, applies to:
    • Amazon Simple Storage Service (S3)
    • Rackspace Cloud Files (Mosso)
    • Ninefold Cloud Storage
    • Dropbox Web Service
    • Remote Servers (Only Protocols: FTP, SFTP, SCP)
    • Local Storage
  • Incremental Backups, applies to:
    • Remote Servers (Only Protocols: RSync)

Cycling Wiki Page

Splitter Wiki Page


  • RSync (Push, Pull and Local)
  • Amazon S3
  • Rackspce Cloud Files

Syncer Wiki Page


  • Gzip
  • Bzip2
  • Pbzip2
  • Lzma

Compressors Wiki Page


  • OpenSSL
  • GPG

Encryptors Wiki Page


  • Shasum (not working with splitter)


  • Mail
  • Twitter
  • Campfire
  • Presently
  • Prowl
  • Hipchat

Notifiers Wiki Page

Supported Ruby versions (Tested with RSpec)

  • Ruby 1.9.3
  • Ruby 1.9.2
  • Ruby 1.8.7

A sample Backup configuration file

This is a Backup configuration file. Check it out and read the explanation below. Backup has a great wiki which explains each component of Backup in detail., 'A sample backup configuration') do

  split_into_chunks_of 4000

  database MySQL do |database|               = 'my_sample_mysql_db'
    database.username           = 'my_username'
    database.password           = 'my_password'
    database.skip_tables        = ['logs']
    database.additional_options = ['--single-transaction', '--quick']

  database MongoDB do |database|             = 'my_sample_mongo_db'
    database.only_collections = ['users', 'events', 'posts']

  archive :user_avatars do |archive|
    archive.add '/var/apps/my_sample_app/public/avatars'

  archive :logs do |archive|
    archive.add     '/var/apps/my_sample_app/logs/production.log'
    archive.add     '/var/apps/my_sample_app/logs/newrelic_agent.log'
    archive.add     '/var/apps/my_sample_app/logs/other/'
    archive.exclude '/var/apps/my_sample_app/logs/other/exclude-this.log'

  encrypt_with OpenSSL do |encryption|
    encryption.password = 'my_secret_password'

  compress_with Gzip do |compression| = true

  store_with SFTP, "Server A" do |server|
    server.username = 'my_username'
    server.password = 'secret'
    server.ip       = ''
    server.port     = 22
    server.path     = '~/backups'
    server.keep     = 25

  store_with SFTP, "Server B" do |server|
    server.username = 'my_username'
    server.password = 'secret'
    server.ip       = ''
    server.port     = 22
    server.path     = '~/backups'
    server.keep     = 25

  store_with S3 do |s3|
    s3.access_key_id      = 'my_access_key_id'
    s3.secret_access_key  = 'my_secret_access_key'
    s3.region             = 'us-east-1'
    s3.bucket             = 'my_bucket/backups'
    s3.keep               = 20

  sync_with S3 do |s3|
    s3.access_key_id     = "my_access_key_id"
    s3.secret_access_key = "my_secret_access_key"
    s3.bucket            = "my-bucket"
    s3.path              = "/backups"
    s3.mirror            = true

    s3.directories do |directory|
      directory.add "/var/apps/my_app/public/videos"
      directory.add "/var/apps/my_app/public/music"

  notify_by Mail do |mail|
    mail.on_success = false
    mail.on_warning = true
    mail.on_failure = true

  notify_by Twitter do |tweet|
    tweet.on_success = true
    tweet.on_warning = true
    tweet.on_failure = true


Brief explanation for the above example configuration

First, it will dump the two Databases (MySQL and MongoDB). The MySQL dump will be piped through the Gzip Compressor into sample_backup/databases/MySQL/my_sample_mysql_db.sql.gz. The MongoDB dump will be dumped into sample_backup/databases/MongoDB/, which will then be packaged into sample_backup/databases/MongoDB-#####.tar.gz (##### will be a simple unique identifier, in case multiple dumps are performed.) Next, it will create two tar Archives (user_avatars and logs). Each will be piped through the Gzip Compressor into sample_backup/archives/ as user_archives.tar.gz and logs.tar.gz. Finally, the sample_backup directory will be packaged into an uncompressed tar archive, which will be piped through the OpenSSL Encryptor to encrypt this final package into YYYY-MM-DD-hh-mm-ss.sample_backup.tar.enc. This final encrypted archive will then be transfered to your Amazon S3 account. If all goes well, and no exceptions are raised, you'll be notified via the Twitter notifier that the backup succeeded. If any warnings were issued or there was an exception raised during the backup process, then you'd receive an email in your inbox containing detailed exception information, as well as receive a simple Twitter message that something went wrong.

Aside of S3, we have also defined two SFTP storage methods, and given them two unique identifiers Server A and Server B to distinguish between the two. With these in place, a copy of the backup will now also be stored on two separate servers: and

As you can see, you can freely mix and match archives, databases, compressors, encryptors, storages and notifiers for your backups. You could even specify 4 storage locations if you wanted: Amazon S3, Rackspace Cloud Files, Ninefold and Dropbox, it'd then store your packaged backup to 4 separate locations for high redundancy.

Also, notice the split_into_chunks_of 4000 at the top of the configuration. This tells Backup to split any backups that exceed in 4000 MEGABYTES of size in to multiple smaller chunks. Assuming your backup file is 12000 MEGABYTES (12GB) in size, then Backup will take the output which was piped from tar into the OpenSSL Compressor and additionally pipe that output through the split utility, which will result in 3 chunks of 4000 MEGABYTES with additional file extensions of -aa, -ab and -ac. These files will then be individually transfered. This is useful for when you are using Amazon S3, Rackspace Cloud Files, or other 3rd party storage services which limit you to "5GB per file" uploads. So with this, the backup file size is no longer a constraint.

Additionally we have also defined a S3 Syncer ( sync_with S3 ), which does not follow the above process of archiving/compression/encryption, but instead will directly sync the whole videos and music folder structures from your machine to your Amazon S3 account. (very efficient and cost-effective since it will only transfer files that were added/changed. Additionally, since we flagged it to 'mirror', it'll also remove files from S3 that no longer exist). If you simply wanted to sync to a separate backup server that you own, you could also use the RSync syncer for even more efficient backups that only transfer the bytes of each file that changed.

There are more archives, databases, compressors, encryptors, storages and notifiers than displayed in the example, all available components are listed at the top of this README, as well as in the Wiki with more detailed information.

Running the example

Notice the, 'A sample backup configuration') do at the top of the above example. The :sample_backup is called the trigger. This is used to identify the backup procedure/file and initialize it.

backup perform -t [--trigger] sample_backup

Now it'll run the backup, it's as simple as that.

Automatic backups

Since Backup is an easy-to-use command line utility, you should write a crontask to invoke it periodically. I recommend using Whenever to manage your crontab. It'll allow you to write to the crontab using pure Ruby, and it provides an elegant DSL to do so. Here's an example:

every 6.hours do
  command "backup perform --trigger sample_backup"

With this in place, run whenever --update-crontab backup to write the equivalent of the above Ruby syntax to the crontab in cron-syntax. Cron will now invoke backup perform --trigger sample_backup every 6 hours. Check out the Whenever project page for more information.


See the Wiki Pages.

Suggestions, Bugs, Requests, Questions

View the issue log and post them there.


Contributor Contribution
Brian D. Burns ( burns ) Core Contributor
Aditya Sanghi ( asanghi ) Twitter Notifier, Dropbox Timeout Configuration
Phil Cohen ( phlipper ) Exclude Option for Archives
Arun Agrawal ( arunagw ) Campfire notifier
Stefan Zimmermann ( szimmermann ) Enabling package/archive (tar utility) support for more Linux distro's (FreeBSD, etc)
Mark Nyon ( trystant ) Helping discuss MongoDump Lock/FSync problem
Bernard Potocki ( imanel ) Helping discuss MongoDump Lock/FSync problem + Submitting a patch
Tomasz Stachewicz ( tomash ) Helping discuss MongoDump Lock/FSync problem + Submitting a patch
Paul Strong ( lapluviosilla ) Helping discuss MongoDump Lock/FSync problem
Ryan ( rgnitz ) Helping discuss MongoDump Lock/FSync problem
Robert Speicher ( tsigo ) Adding the --quiet [-q] feature to Backup to silence console logging
Jon Whitcraft ( jwhitcraft ) Adding the ability to add additional options to the S3Syncer
Benoit Garret ( bgarret ) Presently notifier
Lleïr Borràs Metje ( lleirborras ) Lzma Compressor
Jonathan Lassoff ( jof ) Bugfixes and more secure GPG storage
Michal Cichra ( mikz ) Wildcard Triggers
Dmitry Novotochinov ( trybeee ) Dropbox Storage
Emerson Lackey ( Emerson ) Local RSync Storage
digilord OpenSSL Verify Mode for Mail Notifier
stemps FTP Passive Mode
David Kowis ( dkowis ) Fixed PostgreSQL Password issues
Jonathan Otto ( jotto ) Allow for running PostgreSQL as another UNIX user
João Vitor ( joaovitor ) Changed default PostgreSQL example options to appropriate ones
Manuel Alabor ( swissmanu ) Prowl Notifier
Joseph Crim ( josephcrim ) Riak Database, exit() suggestions
Jamie van Dyke ( fearoffish ) POpen4 implementation
Harry Marr ( hmarr ) Auth URL for Rackspace Cloud Files Storage
Manuel Meurer ( manuelmeurer ) Ensure the storage file (YAML dump) has content before reading it
Jesse Dearing ( jessedearing ) Hipchat Notifier
Szymon ( szymonpk ) Pbzip2 compressor

Want to contribute?

  • Fork/Clone the develop branch
  • Write RSpec tests, and test against:
    • Ruby 1.9.3
    • Ruby 1.9.2
    • Ruby 1.8.7
  • Try to keep the overall structure / design of the gem the same

I can't guarantee I'll pull every pull request. Also, I may accept your pull request and drastically change parts to improve readability/maintainability. Feel free to discuss about improvements, new functionality/features in the issue log before contributing if you need/want more information.

Easily run tests against all three Ruby versions

Install RVM and use it to install Ruby 1.9.3, 1.9.2 and 1.8.7.

rvm get latest && rvm reload
rvm install 1.9.3 && rvm install 1.9.2 && rvm install 1.8.7

Once these are installed, go ahead and install all the necessary dependencies.

cd backup
rvm use 1.9.3 && gem install bundler && bundle install
rvm use 1.9.2 && gem install bundler && bundle install
rvm use 1.8.7 && gem install bundler && bundle install

The Backup gem uses Guard along with Guard::RSpec to quickly and easily test Backup's code against all four Rubies. If you've done the above, all you have to do is run:

bundle exec guard

from Backup's root and that's it. It'll now test against all Ruby versions each time you adjust a file in the lib or spec directories.

Or contribute by writing blogs/tutorials/use cases