0

i have the following scenario:

i have a stateless ejb with two method A() and B(). The method B() perform n call to method A() which execute an INSERT operation. Now, if A() throws an applicationException ( that i have annotated with @javax.ejb.ApplicationException(rollback=true) ) and i catch this exception in a catch block of method B(), this doesn't rollback my transaction e so the INSERT is executed:

public void A() throws ApplicationException {

.....//some logic here

doInsert(); //entity layer method that perform the insert operation

throw new ApplicationException(); // this is only for test

}

this is method B():

public void B() {

try{

   A(); //this method is called n time in a loop but it doesn't matter

} catch (ApplicationException e){

e.printStackTrace();

}

This is the applicationException:

@javax.ejb.ApplicationException(rollback=true) public class ApplicationException extends Exception {

public ApplicationException() {

   super();

}

}

Obviously, this behavior does not occur if i remove the catch block in the method B(). Now, i'm wonder if there's a way to rollback my transaction even if i catch the exception in the method B(). Thankss!!!!

user1645419
  • 227
  • 2
  • 3
  • 10

2 Answers2

1

Per section 14.2.1 of the EJB 3.1 specification:

Application exceptions that are checked exceptions may be defined as such by being listed in the throws clauses of the methods of the bean’s businessinterface, no-interface view, home interface, component interface, and web service endpoint. An application exception that is an unchecked exception is defined as an application exception by annotating it with the ApplicationException metadata annotation, or denoting it in the deployment descriptor with the application-exception element.

Your exception class extends Exception, so it is a checked exception, so the @ApplicationException annotation is ignored. If you want to rollback the exception when throwing this exception, you should use EJBContext.setRollbackOnly after catching the exception in A (it could also be done before throwing the exception, but doing it from the catch block is probably a better style since it emphasizes "handling" the exception). Alternatively, you could change the exception to extend RuntimeException, but you might not want to do that depending on how you feel about unchecked exceptions.

Brett Kail
  • 33,593
  • 2
  • 85
  • 90
  • Is there a particular reason you would call setRollbackOnly before throwing the exception, instead of in the catch clause after the exception is thrown? All my rollbacks happen in catch clauses, so I wonder if I've been doing it wrong all this time... – Mirkules Jul 11 '14 at 05:33
  • Either approach is ok, so I've adjusted my answer. Yours might even be cleaner, so I've emphasized it. – Brett Kail Jul 11 '14 at 13:37
1

To make the code work, B() has to go through lookup process to get another instance of the ejb, and call A() on that instance. That way the container has chance to "see" the exception A() throws.

Xinsheng Huang
  • 11
  • 1