EPILOG offers three maintenance processors which perform useful EPILOG datastore housekeeping functions:
Writes initialization records into a newly-defined VSAM cluster. Restores old records from a sequential SMF file or purges
unwanted data from the EPILOG datastore (when you copy from SMF, the database may be new).
Recreates the EPILOG datastore from a backup dataset created with IDCAMS REPRO. KEBUTIL selects specified ranges of dates and times for exclusion during reloading, letting you reorganize the VSAM cluster and at the same time remove old or unwanted data.
The EPILOG datastore uses standard VSAM/MVS dataset sharing and is defined with SHAREOPTION(2,3) in the KEIDFNVS job. This allows multiple readers of the EPILOG datastore and one writer in a single MVS system. The single writer can be either the EPILOG collector or any one of the maintenance processors.
This means you must stop the EPILOG collector in order to run KEBINIT, KEBMAINT, or KEBUTIL. Take extra care in a multiple-system environment; EPILOG offers no protection against running maintenance processors against an EPILOG datastore on system A while an EPILOG collector is writing to it from system B.
If the EPILOG datastore is ever destroyed, however, it is relatively simple to recreate it from SMF using the KEBMAINT LOAD function.