Q: How do I get jboss to recognise the data_dir

A: All you have to do is to tell jboss to use your own copy of the data_dir instead of the compiled in version.  To do this you have to define an environment variable called JAVA_OPTS.  The setup for this operating system specfic.

Linux/Unix/MacOSX

JAVA_OPTS="-DGEOSERVER_DATA_DIR=/home/user/geoserver-config"
export JAVA_OPTS 

Windows

set JAVA_OPS"-DGEOSERVER_DATA_DIR=C:\home\user\geoserver-config"