Recovery From Errors in Shared Resources
Original Publication Date: 1985-Sep-01
Included in the Prior Art Database: 2005-Feb-19
Concurrently, operating processes sharing a resource, such as a random- access dynamic memory, can be affected in various degrees by an error in the shared resource. The error, usually a hardware error, stops all processes. An evaluation of the propagation of the error is made as to the processes not affected by the error. They are automatically restarted from the point of stopping, while all other processes having exposure to the error are retried using known retry procedures. The recovery procedure is useful in diverse apparatus including a buffer memory which allows loading or receiving signals from a first source while simultaneously supplying signals to a data synch. Examples of such buffers include caches attached to backing storage de vices such as disk files, tape units and the like.