Originally posted by Jaikiran Pai:
I guess, yes.
OK. I'm guessing then, that this isn't something which is easily done, or at least, not done often. Perhaps there is a reason for that?
There are two threads in question - one is involved in storing database text in Lucene indexes and the other is an RMI
thread that allows Lucene searches across multiple boxes (a distributed search). The first thread periodically connects to the database, checks for new rows in one table, and indexes any new text it finds with into a directory of Lucene indexes.
I'm wondering if we shouldn't be looking at a completely different approach. Perhaps we should have a multithreaded "server" (the RMI thread, and the Lucene thread) running on each box, and only 1 JBoss box handling HTTP requests only?
Jon