Integration

Hi all,

I have a BPEL process with a catchall exception block for the entire process (in the .bpel file). The errorhandling determines if action needs to be taken on a given exception and notify's tech support with an e-mail if action needs to be taken. This works fine.
However, it would be nice if you could retry faulted instances if for instance a web service was temporarily down. Here the fault management framework comes to the rescue. I have created default policy and fault-binding files. I have set it up so that any error goes to human interaction. Apparently this worked. An instance pending manual recovery was created, and it could be retry'ed from the BPEL console.

So here is my problem: The fault management framework apparently overwrites the catchall exception in the BPEL process, so that while a task for human interaction is created the error block is not excuted first - meaning that no e-mail notification is sent and no differentiating between exception occurs.
If I remove the default policy and fault-binding files then the catchall is used again.

This may be expected behaviour, but I assume that it is possible to use both a catch all block in the bpel code AND the generic fault management framework.
Does anybody have any ideas?

Your assistance will be much appreciated.
Regards,
Aagaard