I've read the tutorial about how to work in a clustered environment, specially the part that has to be with sequencers.
Using a Lock Column is a good solution, but as the nextVal cannot be rolled back, we would miss many possible codes.
So I have implemented a sequencer that extends AbstractSequenceManager and implements getUniqueLong() executing a SELECT that returns the MAX(ID) from the table. To be sure that it works in a clustered environment, I assure that the broker that is passed to the constructor of my sequencer has set transaction isolation level to REPEATABLE_READ, as EJB do.
I would appreciate some suggestions and posible alternatives or misconceptions I may have with this solution.
Thanks in advance, Enrique Medina.
_________________________________________________________________
�Cu�nto vale tu auto? Tips para mantener tu carro. �De todo en MSN Latino Autos! http://latino.msn.com/autos/
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
