I have a doubt regarding the behaviour of a Scheduled Bean that must be deployed over a clustered environment.
1 - My guess is that it will be deployed over all the instances of the cluster (workers) and by doing that I would have 4 scheduled beans running, one in eacho worker of the cluster since I have a cluster with 4 nodes. Is that so?
2 - My scheduled bean checks a table, doing a pooling behaviour. The project that this bean belongs uses JPA + Hibernate. If I have 4 schedulers running each one inside a Context (that is related to a cluster node or worker) I could suffer with repeatable readers and waste processing on a task that is already beeing processed by the scheduler of another node. This situation is addressed how? Does JBoss hadles this automatically or should i take a deeper concern about that using a transaction isolation level more restrict?