On Sat, Mar 7, 2009 at 2:46 PM, Angelo Immediata <[email protected]> wrote: > In my project i need to handle a digitla repository very great (around 4-5 > TB). Is able Jackrabbit to handle this kind of repository? > Are there any problems?
Jackrabbit in fact is optimized for storing large binaries through the use of the DataStore [1], where the file based implementation simply stores the binaries in the filesystem and streams it more-or-less directly from the client to the hard disk, so that you get as much as speed as possible. It also ensures that two identical binaries are only stored once in the datastore, even if they occur multiple times in the JCR tree. Indexing of large binaries (via text extractors) is one factor that can take some time (but that depends whether those files can be indexed at all, can be configured), and I personally haven't seen such a large repository based on Jackrabbit. But I can't see any general problems. [1] http://wiki.apache.org/jackrabbit/DataStore Regards, Alex -- Alexander Klimetschek [email protected]
