Optional: Customizing the randomizer for exceptionally large database records

Tuning areas that contain exceptionally large database records can be difficult.

Before you begin

In this task, you will move large database records to new areas to simplify the tuning tasks. Before you start, review the applications that use the database and ensure that those applications are not affected by moving such records to the new areas.

About this task

If your administration policies allow, one possible approach to tuning areas that contain exceptionally large database records is to define new areas that have enough CI size, and to move these large records to the new areas before tuning the areas.

Procedure

Complete the following steps to define new areas and to move large database records to the new areas:

  1. Determine whether you can move large database records to different areas. Then obtain the information about the large database records to move.

    Generate a Largest Database Records report and browse the report. This report provides a list of database records and the information about these records. Analyze the information in the report to determine whether you can move those records or a subset of those records to a new area that matches the record characteristics.

  2. Customize the randomizing module for the selected database records and determine the DBD statements for the new areas.

    You must ensure that the selected large database records be randomized into the new areas, which means that you must customize the randomizing routine for the database. In general, the randomizing routine that is used needs to be a two-stage randomizer. In this context, a two-stage randomizer determines an area from the root sequence key value and then determines a RAP number within the selected area.

    To customize or tune the randomizing routine, you need at least the root sequence key values for the selected database records and information about the database record length for each key. You can obtain this information from the Largest Database Records report.

    When you customize the randomizing routine, be aware that the root segments of those large database records can be randomized on different RAPs in new areas.

    For the information that you can obtain from the Largest Database Records report, see Step 5 of the topic Analyzing database record profile.

What to do next

Follow the steps in Tuning randomizer and DBD parameters for DEDB to tune the areas.