As I've got a DMI call, it's something that will run across the cluster.
But is it allow to throw an exception within a DMI method.
node1 proceed well,
node2 throws an exception Ex1.
node3 throws an exception Ex2.
What will be the result for the callee thread. How to catch Ex1 and Ex2?
Is there something that may help to rollback the work on the node1?
Easy way is to not allow exceptions. But what to do if you've got a failure on nodeX?
Or on all nodes?
I want to run something over all nodes that uses local resources. And that something has rights to fail.
I'm looking for a way to notify the callee thread about a failure.
Command pattern you refer to is just a java.lang.Runnable. Thus there's no way to throw an exception.
I try to understand what you mean but cant grasp your thoughts.