System and Method for generalized HACMP/failure safe handling of entity beans
Original Publication Date: 2005-Mar-25
Included in the Prior Art Database: 2005-Mar-25
This innovation describes a system and method for generalized support for safe handling of failure exceptions resulting from stale database connections in a Java* J2EE* entity bean environment such that specialized retry code for each entity bean is not necessary.
System and Method for generalized HACMP /failure safe handling of entity beans
When writing proper J2EE code that utilizes an entity bean, there is a problem in that it is possible for the database connection layer to throw a stale connection exception as a result of a database connection failure. This can happen if the database fails, and temporarily is unavailable during a switch-over to a backup database. The connections in the connection pool that point to the faulty database are now considered stale.
Unfortunately, application software is required to handle this case. The problem is that this requires each entity bean operation to be wrapped with special exception handling code that retries a number of times in an attempt to get a valid connection to the new database. This code typically requires a specific reference to the entity bean, of which there may be many in the system, and additionally many programmers do not realize that they need to put this special handling code throughout their software, leaving them vulnerable to a recovered failure on the database resulting in an application failure.
The solution is to provide an easy to use utility mechanism to generically manage the entity beans and provide a layer between the application business logic and the entity bean, such that the automatic retry logic during a failure can be done in a consistent and comprehensive manner.
This solution enables dynamic introspection and invocation of entity beans such...