A simple failover between nodes with a little load; the setup is identical to the one described here, i.e. single box, three EAP instances acting as workers and one acting as a balancer. A session is established with one of the workers and the one is killed shortly after that. Balancer forwards the next request to a new worker.

• See ERROR org.infinispan.util.concurrent.TimeoutException
• See ISPN000141: Could not rollback prepared 1PC transaction.
• See ISPN000136: Error executing command PrepareCommand, writing keys []