Exception Handling with OC4J Web Services

Question

Hi,
I want to throw some custom exceptions from my web services, based upon my business logic. From the documents I came to know that I can use "javax.xml.rpc.soap.SOAPFaultException" for the same. Following is the sample web service code which I'm trying in my environment.
import javax.ejb.Stateless;
import javax.jws.WebMethod;
import javax.jws.WebService;
import javax.xml.namespace.QName;
import javax.xml.rpc.soap.SOAPFaultException;
@WebService
@Stateless
public class TestService {
     @WebMethod
     public void greeting()
          throw new SOAPFaultException(new QName("uri", "local"),
     "My Fault String", "My Fault Actor", null);
Deployment of the web service goes fine and on the invocation of the "greeting" operation the exception is being thrown. But the problem is that, the soap response, which my web service client receives is as of follows,
<env:Envelope
xmlns:env="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:ns0="http://service.csm.nb.md.inglife.jp.co/">
<env:Body>
<env:Fault>
<faultcode>env:Server</faultcode>
<faultstring>Internal Server Error (Caught exception while handling request: oracle.oc4j.rmi.OracleRemoteException: javax.xml.rpc.soap.SOAPFaultException: My Fault String; nested exception is: javax.xml.rpc.soap.SOAPFaultException: My Fault String)</faultstring>
</env:Fault>
</env:Body>
</env:Envelope>
This means that OC4J again wraps the SOAPFaultException thrown by me. Is there any way by which I can avoid the further wrapping of my exceptions? If there is anything wrong in my approach please do let me know.
Regards,
Dipu

Answer 1

This is one of the "classic" design problems in this kind of architecture. And, unfortunately, the answer is "it depends on how you think you need to handle it." And I'm sure there are plenty of "gurus" that will tell you one way or another is the only way to do it.
I'll be more honest: I'll give you a couple of personal suggestions, based on experience in this architecture. These are suggestions - you may do with them what you will. I will not say this is the best, most correct, or even remotely relevant to what you're doing.
If it's simple data validation for "typing" (e.g. String, number, Date, etc.), that is taken care of when you attempt to stuff in the information into the appropriate DTO. If it's more "sophisticated" than that (must be in a certain range, etc.), that particular checking should probably be delegated from your Controller to a helper class. That not only saves the "expense" of transmitting the information back and forth across the wire, it's "faster" to the end user so say "Ooopsie" by redirecting back to the form right then. Basically the same thing if the types are wrong.
That only leaves the "big" problems in the business layer (EJBs), where you have to deal with concurrency, database failures, etc. Generally these kinds of exceptions are thrown back to to the Controller in one of two forms:
1) a sublass of RuntimeException, which signals that some Very Bad Things have happened in your container. EJBException is one like that and you can see where it's being thrown from.
2) a subclass of Exception, also called "application exceptions." They are usually something like a "duplicate record" or a validation-like error (which you mentioned) like a missing field. They're used as a signal to a failure in the logic, not the container. That way you have to decide at what layer of your architecture they should be handled and/or passed on to the next.