Guidelines for Recovering a Workflow Process

It is recommended that you adhere to the following guidelines to recover a Workflow Process:

• To reduce the number of checkpoints and minimize runtime overhead, you should modify the Allow Retry Flag property on Siebel operation steps and business service steps. If the Workflow Recovery Manager cannot determine from which step it must recover a Workflow Process instance, then it marks this instance for manual recovery. To recover the instance, the Workflow Recovery Manager examines the state information for the instance that the Workflow Engine saves at recovery checkpoints. To optimize performance, the Workflow Engine determines the recovery checkpoints according to the step type and the Allow Retry Flag property on the step.

• Separate a complex business service into a number of simpler business services. It is easier to recover from a smaller business service.

• To avoid transient or permanent system problems, use multiple Workflow Recovery Managers. For example, a situation can exist in an environment where the subnet in which the Workflow Recovery Manager resides is frequently down. It is typical for only one Workflow Recovery Manager to exist as long as the Siebel Server can restart it if this server fails.