Attempting to access a distributed object from a remote node when the master node for the object is down will cause a com.kabira.platform.ResourceUnavailableException exception to be thrown.

A remote node is detected to be down when the following action is attempted on a distributed reference:

Highly available objects that have replica nodes defined will never throw the com.kabira.platform.ResourceUnavailableException exception because the master node of the distributed object fails over to a replica node when the current active node for the partition fails. This failover is transparent to applications (see the ActiveSpaces® Transactions Architect's Guide).