Is it possibe with the new 4.5 BerkeleyDB to use the Replication
Framwork as follows:

(1) Have a program that updates the database
(2) Have another program that attaches to the same database and
initiates itself as a master replication program.
(3) Have another program residing on another machine that initiates
itself as a slave that replicates the database.

If so can the replication framework be setup so that (2) and (3) are
not always running?

At the moment I simply use rsync to replicate the log files from one
machine to another (each minute) and then run db_recover every hour,
this works but it would be nice to have something more intuitive
without having to implement real-time replication.

Gerard