Data and system flow of DEDB Tuning Aid
The DEDB Tuning Aid utility uses the sorted and resolved data analysis records generated by the DEDB Pointer Checker, and a DMB that defines the database structure to be tuned.
If only the randomizing routine is being changed, the old DMB can be used and the new randomizer name can be supplied by means of a control statement parameter. Otherwise, a new DMB must be provided.
The first step reprocesses the data analysis records and, if required, substitutes the new AREA and RAP numbers. New analysis records are only generated for requested areas.
The second step uses the sorted segment information records to pseudo insert the segments into an area model, and the appropriate freespace, reference, and segment information records are generated.
The final step creates the result of the tuning process: a set of output reports formatted by the DEDB Pointer Checker report generation module.
DEDB Tuning Aid programs run in a virtual storage region as standard z/OS® batch job steps. Region size varies with the size of a UOW and the number of IOVF CIs defined. Typical usage should start at 768K which includes requirements for the DEDB Tuning Aid programs, access methods and buffers, and report queues.
DEDB Tuning Aid contains three load modules. The following table lists these modules.
| Load module name | Function |
|---|---|
| FABBME1 | Extract/format segment data |
| FABBME3 | Pseudo load area, generate records for reporting |
| FABADA5 | Generate, format, and print reports |
Processes of DEDB Tuning Aid
A typical DEDB Tuning Aid job contains five or six of the following steps:
- Step 1. FABBME1
- This FPB program sets up the environment required for invoking a randomizer. It reprocesses the data analysis records (that were generated by the DEDB Pointer Checker) substituting the new AREA and RAP numbers. It then generates analysis records for the requested areas. One output file containing all record types is created.
- Step 2. DFSORT
- The IBM® Data Facility Sort Program Product 5740-SM1 (or its equivalent) sorts the analysis records created by FABBME1 (file MEDARO).
- Step 3. FABBME3
- This FPB program
sets up an area model and pseudo inserts all segments into the model.
It writes the area description, segment information, and free space
records required by the reporting program to two output files:
- The first data set contains area description, segment definition, and free space information records.
- The second data set contains the segment information records.
- Step 4. DFSORT
- The IBM Data Facility Sort Program Product 5740-SM1 (or its equivalent) sorts the area description, segment definition, and free space information records created by FABBME3 (file MEDAR120).
- Step 5. DFSORT
- If more than one area is being processed, the IBM Data Facility Sort Program Product 5740-SM1 (or its equivalent) sorts the segment information records created by FABBME3 (file MEDAR30). Omit this step if only one area is requested. This is the only optional step.
- Step 6. FABADA5
- This FPB program uses the sorted files from the previous two steps to generate and print the analysis reports. Optionally, threshold analysis is performed, reorganization control statements are generated, and a history file is written.
The following figure shows the flow of the DEDB Tuning Aid procedure.
