rake_for_postgres_s3_backup

#First you have add the following gems to your Gemfile
gem 'aws-sdk', '< 2.0'
gem 'zip'
gem 'whenever'


#Created a new file at lib/tasks/backup.rake and add the following
desc "PG Backup"
namespace :pg do
  task :backup => [:environment] do
    #stamp the filename
    datestamp = Time.now.strftime("%Y-%m-%d_%H-%M-%S")

    #drop it in the db/backups directory temporarily
    app_name = 'marketplace-ui'
    backup_file = "#{Dir.home}/db_backups/#{app_name}_#{Rails.env}_#{datestamp}_dump.sql.gz"

    #dump the backup and zip it up
    user = Rails.application.secrets.pg_user
    password = Rails.application.secrets.pg_password
    database = Rails.application.secrets.database
    sh_cmd = "PGPASSWORD=#{password} pg_dump -h localhost -U #{user} #{database}| gzip -c > #{backup_file}"
    sh sh_cmd
    send_to_amazon backup_file
    #remove the file on completion so we don't clog up our app
    File.delete backup_file
  end

  def send_to_amazon(file_path)
    s3_config = YAML.load_file Rails.root.join("config/aws.yml")
    file_name = File.basename(file_path)
    s3 = AWS::S3.new(:access_key_id => s3_config[Rails.env]["access_key_id"],:secret_access_key => s3_config[Rails.env]["secret_access_key"])
    #push the file up
    bucket = s3.buckets[s3_config[Rails.env]["bucket"]+"/db_backups"]
    bucket.objects.create(file_name, File.open("#{file_path}"))
  end
end


#Add the following line to config/schedule.rb 

every :day, :at => '12:30am' do
  rake "pg:backup"
end 


#Add the following line to config/deploy.rb 
set :whenever_command, "bundle exec whenever"
require 'whenever/capistrano'


Note: This Gist is based on this old post
http://blog.biasedwalk.com/2013/07/backing-up-postgres-to-amazon-s3-with.html?showComment=1450286410055
but the above author uses aws-sdk version 1, here we are using aws-sdk version 2