Okay, we having a little debate here about whether or not the following statements from the weblogic 5.1 documentation are to be interpreted "literally".

"When using WebLogic JNDI to replicate bindings, the bound object will be handled as if it owned by the host server -- the server that initially bound the object to it's JNDI tree. If the host server fails, the object will be removed from all the servers in the cluster. While this is the appropriate behavior in cases where the bound object is an RMI stub referring back to the host server, it may not be the appropriate behavior for your object. This can have an adverse effect on the availability of your object. If the availability of your object is critical, you can use an approach similar to the one described in Data caching example. "

So if we're thinking about maintaining data in the JNDI across the nodes in the cluster and the server that inserted the data crashes are these data wiped out in the cluster or not. I take the foregoing WL documenatation according to the literal view. Others think it applies only to RMI stub replication?

Any input would be appreciated....