Performance improvement

IMS Database Recovery Facility should provide an improvement over IMS database recovery procedures for the time that is required to recover database data sets.

This improvement is due to the following reasons:

  • Log data sets that must be read to recover the set of database data sets are read in parallel.
  • Multiple database data sets and areas are recovered simultaneously.
  • Shared database data sets and areas do not require a change accumulation step before running the IMS Database Recovery Facility.

    Not running change accumulation before recovering shared databases can be beneficial to database recovery performance.

  • Multiple processing steps are not required when using auxiliary utilities.

    Each database is read only once for processing by IMS Database Recovery Facility, along with IMS High Performance Pointer Checker and IMS High Performance Image Copy (if IMS High Performance Pointer Checker or IMS High Performance Image Copy was requested in batch).

    Processing by IMS Index Builder, the DFSPREC0 utility, or the Build Index function of FPA is not included.

    After the database is processed by IMS Database Recovery Facility, it is passed to two separate tasks, one for IMS High Performance Image Copy and one for IMS High Performance Pointer Checker, which run in parallel. IMS Database Recovery Facility continues processing as a third task.

    Before IMS Database Recovery Facility reads the next database, it waits for completion of the other two tasks.

    Using this method improves performance because databases need to be read only once, not three times, and parallel processing occurs.

Performance of each recovery depends on your processing environment at the time of the recovery.

The improvement in performance depends on the following:
  • The number of log data sets that are required for recovery
  • The number of tape drives and physical data paths that are available to IMS Database Recovery Facility for reading data sets
  • The number of database data sets and areas that are being recovered