Suspend Message Queue
Original Publication Date: 1981-Nov-01
Included in the Prior Art Database: 2005-Feb-12
This article describes a method for operating a plurality of subsystems sharing access to a common data base to isolate failures in one subsystem from another non-failing subsystem. Requests for locks on a data base portion, rejected because a lock on that portion had been previously granted to a subsystem which subsequently failed without releasing its locks, are held in a suspend message queue, and released for reprocessing only after the failed subsystem has been restarted.