Replication of a file from the default US region into an Asia bucket.

file_replication_multi-region.rb

#!/usr/bin/env ruby

require 'rubygems'
require 'fileutils'
require 'date'
require 'chronic'
require 'fog'

#Set the default directory for relative paths.
Dir.chdir(File.expand_path(File.dirname(__FILE__)))

# Import EC2 credentials e.g. @aws_access_key_id and @aws_access_key_id
require './s3_credentials.rb'

## S3 Upload
# Create a connection
connection = Fog::Storage.new({
  :provider              => 'AWS',
  :aws_access_key_id     => @aws_access_key_id,
  :aws_secret_access_key => @aws_secret_access_key
})

asia_connection = Fog::Storage.new({
  :provider              => 'AWS',
  :aws_access_key_id     => @aws_access_key_id,
  :aws_secret_access_key => @aws_secret_access_key
  :region                => "ap-southeast-1"
})

new_file = connection.copy_object("csi-testbucket-eric", "calgary.pdf","csi-testbucket-eric-asia","calgary.pdf", {'x-amz-acl' => 'public-read'})
new_new_file = asia_connection.directories.get("csi-testbucket-eric-asia").files.get("calgary.pdf")

puts new_new_file.public_url


OUTPUT
[WARNING] fog: followed redirect to csi-testbucket-eric-asia.s3-ap-southeast-1.amazonaws.com, connecting to the matching region will be more performant

This one is tricky for sure. The first optimization/simplification is that copy_object is run against the target (not the source). So you should actually be able to remove the first connection part altogether and use the asia_connection throughout. This should remove the warning and avoid the redirect which should make things run a bit faster.

You can optimize #31 by assuming that the bucket and file exist (get will actually look them up, new assumes) like so:

new_new_file = asia_connection.directories.new(:key => 'csi-testbucket-eric-asia').files.new(:key => 'calgary.pdf')

Public url should then still work as expected, but run faster.

Hope that helps explain, both are pretty fine points but they should have some impact.