Removing the use of file system

[Karthik Sudarshan <Ka...@Sun.COM>]

Hi all,
    I have a problem where Jackrabbit repository may be on one machine 
but the appserver (using Deployment Model 2 with jackrabbit-jca rar) may 
be on a different machine. The file system may not necessarily be 
shared, and due to this I wanted to switch to the Derby database 
entirely. Although it was successful for the most part, the searchIndex 
still expects a mandatory path attribute! This is the only thing which 
prevents me from having the whole repository on the DB. Is there 
anything that can be done so that the search indexes are moved to the DB 
as well?

Regards,
Karthik

Re: Removing the use of file system

[Marcel Reutegger <ma...@gmx.net>]

Karthik Sudarshan wrote:
> Hi all,
>    I have a problem where Jackrabbit repository may be on one machine 
> but the appserver (using Deployment Model 2 with jackrabbit-jca rar) may 
> be on a different machine. The file system may not necessarily be 
> shared, and due to this I wanted to switch to the Derby database 
> entirely. Although it was successful for the most part, the searchIndex 
> still expects a mandatory path attribute! This is the only thing which 
> prevents me from having the whole repository on the DB. Is there 
> anything that can be done so that the search indexes are moved to the DB 
> as well?

the current implementation is based on a local file system and cannot be moved 
to a database.

in early days there was support for the FileSystem abstraction but because of 
heavy random access to the files the performance suffered. we then decided to 
use the local file system only.

if index files are stored in a DB they probably have to be spooled to a temp 
file where random access is possible. this however requires additional coding.

regards
  marcel