RECYCLE Utility › Multisystem Considerations › Shared Data Set Hardware Reserve

Shared Data Set Hardware Reserve

When the GROUPENQ DD is coded in the RECYCLE JCL, flags in the GROUPENQ data set are serially accessed under a hardware reserve to determine which groups are being processed. This type of serialization can safely operate in a multisystem environment without IBM's GRS or CA MIM.

Note: The GROUPENQ data set must be placed on a shared DASD volume that is accessible to all of the participating systems.

A RECYCLE job issues a hardware reserve on the shared DASD volume and enqueues the GROUPENQ data set, a flag representing a given group is placed in the data set, and the reserve and enqueue are released. This allows all RECYCLE jobs to determine if a given group is currently being processed, but only one job at a time can access the data set. When a RECYCLE job completes processing a group, the hardware reserve is issued on the shared DASD volume and the GROUPENQ data set is enqueued, the flag is removed from the file, and the reserve and enqueue are released.

This type of serialization can provide performance advantages over an IBM GRS or CA MIM environment as long as the shared DASD volume does not have any significant contention.

If RECYCLE is canceled after it begins processing, the flag representing the current group may not be cleared. Subsequent RECYCLE jobs will not attempt to process a group whose flag has not been cleared. You can manually clear a flag left on in error by deleting and then redefining the GROUPENQ data set.

Sample JCL to allocate and initialize the GROUPENQ data set and to reset a group flag can be found in HLQ.CCUUJCL(RECYRSET).