MDB Failure Handling
Internal
Relevance
EAP 6.4.10
Overview
This article addresses failure handling in an MDB context. It was written while experimenting with EAP 6.4 and a HornetQ-based messaging subsystem.
Failure Handling Specification
JSR 318 Enterprise JavaBeans Version 3.1 EJB Core Contract and Requirements, Section 5.4.18 "Dealing with Exceptions" mentions that MDBs should not throw RuntimeExceptions. If a RuntimeExceptions occurs, the container will transition the MDB in the "does not exist" state. The message will not be acknowledged, and if messages arrive to the destination, the container can delegate the message to another MDB instance.
WildFly/HornetQ Behavior on RuntimeException
Message Delivery Occurs in a Container-Managed Transactional Context
If the message delivery occurs in a container-managed transactional context, the HornetQ resource adapter acknowledges the message when it first receives from the HornetQ JMS provider. This behavior can be thought as "removing the message from the queue". The resource adapter attempts to deliver the message in a loop until either it is consumed by an MDB instance or the message is sent to the DLQ. The number of redelivery attempts is controlled by the dLQMaxResent attribute of the MDB container configuration.
Who loops and Who sends the message to DLQ?
Does this mean that if the JVM fails while the message is processed in the loop, the message is lost?
The current transaction is automatically rolled back. The MDB container's transactional interceptor catches the RuntimeException and rolls back the transaction, by invoking javax.transaction.Transaction.setRollbackOnly().
The MDB instance that triggered the failure is destroyed. The @PreDestroy callback, if exists, is invoked, and the instance is discarded from the pool. Upon re-delivery, if any, a new instance to handle the message will be created.