We are currently using HAJNDI and are wondering if there were any data integrity issues that were possible as data is replicated across a cluster.
For example, if an object is retrieved from HAJNDI, modified, and rebound, does the rebind return before or after the data has been replicated across the cluster? If a new bind into HAJNDI occurs, does the bind call block until the data has been replicated, or immediately after it has been bound locally?
Is there a configuration setting that allows this to behavior to be altered to a synchronous vs. asynchronous call?
This question is also posted in Clusters on JBoss (Clusters/JBoss).