FABBME1 MEDARI DD data set
The FABBME1 MEDARI DD data set is the primary input to the DEDB Tuning Aid. It is the sorted and resolved segment-information-records data set from Step SORTR13 of the DEDB Pointer Checker procedure.
The detail of the procedure is shown in Figure 2. For convenience, these records are also referred to as SRDA13 data.
DEDB Pointer Checker considerations
When the DEDB Tuning Aid is to be used, you must retain the SRDA13 data from the DEDB Pointer Checker run(s) for the area(s) involved. You must also use TYPRUN=MODEL on your DEDB Pointer Checker FABADA1 control statement.
Randomizing routine considerations
The number of areas for which SRDA13 data is required depends on the database specification being changed and the characteristics of the randomizing routine.
For purposes of this document, randomizers are classified as follows:
- Area-specific: Defines the user-written randomizer that controls
an area or area group; its selection is based on root segment key
ranges or a similar technique.
EXAMPLE: Keys starting with "A" through "L" are randomized across areas 1 through 5. Keys starting with "M" through "Z" are randomized across areas 6 through 10.
- Nonspecific: Defines a randomizer where the area selection is not controlled by key values. DBFHDC40, as delivered with IMS Fast Path, is nonspecific.
- The randomizer module interface environment created by FABBME1 conforms to the published interface as defined in IMS Exit Routines with one exception. When a randomizer is invoked by IMS, registers 10 and 11 contain the addresses of the EPST and ESCD, respectively. Some user-written or user-customized randomizers have been designed to use these addresses to gain access to IMS control blocks other than those passed as part of the published interface. When FABBME1 issues a call to a randomizing module, registers 10 and 11 are set to -1 and 0, respectively, to specify that the call is not being issued in a live IMS environment.
- The randomizer module must be capable of AMODE=31.
The areas for which you must provide SRDA13 data are a function of the kind of changes that you want to model. The following table shows how to select the right areas.
Specification changed | Nonspecific randomizer | Area-specific randomizer |
---|---|---|
Randomizing routine | SRDA13 data for all areas | SRDA13 data for the specific area group if the area selection criteria is not changed |
ROOT or UOW values | SRDA13 data for all areas | SRDA13 data for the specific area group |
CI size only | SRDA13 data for the specific area being tuned | SRDA13 data for the specific area being tuned |
Number of areas | SRDA13 data for all areas | SRDA13 data for all areas |
None (that is, tune the Direct Reorganization utility) | SRDA13 data for the specific area being tuned | SRDA13 data for the specific area being tuned |