Fastest way to run IDMSDBAN against multiple areas

Products

• IDMS
• IDMS - Database

Issue/Introduction

Databases may have many dependent areas listed in an UNLOAD.  To run IDMSDBAN against a large area or a large number of areas and sets, may take a very long time. Running one very large IDMSDBAN as a single job may take even longer if the job for any reason needs to be cancelled, or aborts. This article outlines options to make this situation more manageable.

Environment

Release: All supported releases.

Resolution

One of the job steps in running IDMSDBAN executes the program IDMSDBN2. It's not possible to split IDMSDBN2 against one area into smaller jobs, so if any area is large, that job may run long. The overall run-time for IDMSDBAN against a large group of related areas and sets can be reduced by processing fewer sets or areas at any one time. Multiple IDMSDBAN jobs can be run concurrently, each against a different group of areas or sets. The UNLOCKED parameter must be used if there are cross-area sets. 

The following SYSIDMS parameters are recommended:

USERCAT=OFF
PREFETCH=OFF
IDMSQSAM=ON
QSAMAREA=areaname
QSAMBUF#=255
SORTSIZE=ON

Lastly, in the control statements for the SORT package in use, specify the parameter SIZE=E99999999.