Calling s3sync.rb from cron, s3config.yml not getting found

[cleaner416]

First off, thanks for the development of this great tool.

I am trying to automate my syncs via cron, but I'm having some trouble with syntax and getting the script to pickup s3config.yml.  I'm using this command: (Fedora FC5)

/usr/bin/ruby /root/s3sync/s3sync.rb -r --ssl /my_dir_to_backup bucket:pre

Unfortunately, this command only works when I'm in /root/s3sync, and elsewhere errors on the require s3config line (~19).  Is there a way to pass the path of the s3config.yml manually?

[ferrix]

I don't use the yml myself.. I just cd to the s3sync dir before issuing my commands.  Someone familiar with yml maybe could help?

[ledjon]

Just change your process to "cd" to the directory where s3sync.rb (and I assume s3config.yml) is installed at.  For example:

cd /root/s3sync && /usr/bin/ruby ./s3sync.rb -r --ssl /my_dir_to_backup bucket:pre

Since you're fully qualifying the path to backup, the location the script runs from doesn't matter for the process, so this just starts it out in the "installed" directory.

- Jon

[IanTegebo]

I updated s3config.rb to use

Code:

"#{ENV['HOME']}/.s3config.yml"

instead of

Code:

's3config.yml

.

But then you might want to take this a step further and

• move all *.rb files into ~/lib/ruby/1.8/s3sync
• while putting s3{sync,cmd}.rb into ~/bin. 
• Then you need to update each of the libraries to have require s3sync/LIBRARYNAME. 
• Finally, set this environment variable in your login scripts, RUBYLIB=~/lib/ruby/1.8.

[ferrix]

A nice fix for this will be coming today.  Watch announce.