Method for allowing a clustered computer systems manager to use disparate hardware on each of the separate servers utilized for redundancy

Abstract

An apparatus for and method of providing failure recovery from redundancy, notwithstanding that the failed subsystem and its replacement have differing capacities. This is especially useful when implementing a cluster lock processing system having a relatively large number of commodity instruction processors which are managed by a highly scalable, cluster lock manager. Reliability is built into the managing communication processor by dividing the system into master and slave subsystems. The master has primary responsibility for system management and coordination, whereas the slave has primary responsibility to backup the master and be prepared to assume management responsibility. Upon the need to transfer responsibility from the master to the slave, whether it be manual (e.g., maintenance) or automatic (e.g., failure), the only concern is that the slave has sufficient capacity to accept the current level of processing, even though it does not have the same level of capacity as the master.