Messaging System Failure
The messaging system is integral in the workflow system. Any new process is initiated with a ‘workflow’ message.
You can use multiple messaging server instances and set them up in a fault tolerant fashion. If a workflow is suspended and resumed later, the resumption is triggered by a workflow message on the workflow queue. A messaging system failure is deemed to have occurred when the server is not able to connect to the JMS to read or write a message from or in the queue.
Between workflow messages, the workflow typically executes on a single thread in a single instance. Asynchronous activities are an exception to this; the activity workload is split into multiple threads.
The supported messaging solution is EMS. It allows a client aware fault tolerant setup. You can configure a messaging cluster with two members (a primary and a backup server). The addresses of both these servers have to be part of the configuration. If the primary server fails, fallback to the backup server occurs.