Previous | Table of Contents | Next |
The Fault Tolerance Infrastructure includes Logging and Recovery Management Mechanisms that support the infrastructure-controlled
ConsistencyStyle. During normal operation, the Logging Mechanism records the state and actions of the primary member of a
passively replicated object group in a log. After a fault, the Recovery Mechanism retrieves these records from the log and
uses them to restore the state of a backup member of the object group, so that it can continue the service provided by the
primary member that failed. The Logging and Recovery Mechanisms are also used to activate a new member of an actively replicated
object group. No interfaces are defined for the Logging and Recovery Mechanisms because these mechanisms are never invoked
directly by the application program.
This section defines two interfaces that objects of the application program inherit: Checkpointable and Updateable. An application
object that needs to have its state logged and restored must inherit the Checkpointable interface. In addition, it may inherit
the Updateable interface, which allows state changes to be logged and restored incrementally.