In the production environment, the goal is to process all the requests that have entered the system in a methodical and consistent manner. Data preservation is required for this environment and all measures must be taken to minimize system unavailability and data loss.
You need understand which topology type best suits your the production environment. Finding the topology that is right for you requires an analysis of application properties and of nonfunctional requirements.
For more information on topology types, see the topic Planning for a deployment environment based on one of the supplied patterns in the WebSphere® ESB information center.
For example, if a cluster has multiple cluster members, it is possible that the only thing that needs to be recovered is a single-cluster member and that the workload management machinery has already redirected work to ‘running servers'. If this is the case, restarting servers should force recovery and that server should join back into the cluster again.
Some High Availability (HA) configurations have the ability to recover failed transactions from one server on another (known as peer recovery).
Recovery of production data requires success on two levels; system and application.