All in-Memory JCR including Lucene

[Ron Wheeler <rw...@artifact-software.com>]

For a variety of reasons, which perhaps did not take into account the 
learning curve and lack of documentation of Jackrabbit, I thought that 
it would be a good idea to use JCR for an application that runs as a 
batch jobs and then vanishes taking all of its interim data with it.
An in-memory JCR should be almost as fast as a datastructure based on 
Maps and Lists.  Speed is of almost no importance.
It also could be converted to using a disk if the data volumes were too 
large for an in-memory solution by simply changing the repository.xml 
file with no code changes.

I have almost got a fully in memory solution but am still getting stuff 
on the disk.
I believe that the SearchIndex entries are the source of the hard drive 
activity.

I got a working repository.xml file which I have documented in a 
technical note at:
http://blog.artifact-software.com/tech/?p=168

Does it look like a well constructed in-memory configuration?

I was trying to find out the difference between what is configured as 
"FileSystem" as opposed to things configured as the class in the 
persistence manager.
It appeared that discussions that started out with the word FileSystem 
got quickly into talking about persistence.

How do I get rid of the rest of the writing on the hard drive?

Thanks

Ron


-- 
Ron Wheeler
President
Artifact Software Inc
email: rwheeler@artifact-software.com
skype: ronaldmwheeler
phone: 866-970-2435, ext 102