Here is an article on the extraction process.
Information security insights and other ramblings
Searching and extracting data from PST files(in a Linux environment)
http://dereknewton.com/2011/02/searching-and-extracting-data-from-pst-files/
A PST file is an offline message store, typically used for archival of old data or migrating data across domains, and only containing data for one mailbox. What’s on the servers will be EDB files. This is where the current, online data is stored. When you do legal discovery, you plug in search parameters, and the software finds the relevant items in the EDB file, and extracts it out into one or more PST files. Disaster recovery tapes would be backups of the servers.