Fault-tolerant computer system with online reintegration and shutdown/restart

Method of operating a computer system comprises the steps of: executing code by a CPU from memory, including page swapping from said memory and file access to non-volatile storage, in normal operation; detecting a failure of a power supply for said system and initiating a shutdown process in response thereto, said shutdown process including switching to backup power; said shutdown procedure including storing the state of said computer system including the state of processes being executed, in said non-volatile storage; after completing said shutdown procedure, if said power supply has been restored, initiating a restart procedure; said restart procedure including reading said stored state from said non-volatile storage and restarting said processes and continuing execution without rebooting; or, if said power supply has not been restored, shutting down said backup power and ceasing execution by said CPU.