Example 2: Iterative tuning analysis
In this example, an HDAM database (SD148P) and an HIDAM database (SD144P) are being analyzed. The purpose of the analysis is to determine good values for the DBD parameters that pertain to randomization.
Once the analysis is complete, the DBDs will be changed accordingly, and the databases will be reorganized. SD144P will be converted to HDAM organization.
It is assumed that the KEYSIN data set contains root keys for both databases (SD148P and SD144P). The following figure shows the JCL and control statements that must be coded by the user.
//EXAMPLE2 JOB --- use normal job statement parameters here ---
//*
//TUNE1 EXEC FABTIMS,
// CYL=’10,10’, SPACE FOR WORK FILES
// PARM2=’SIZE=(MAX)’, PARM FOR DFSORT
// KEYSIN=’HPS.TEST.KEYSIN’, INPUT KEYS
// DBDLIB=’HPS.TEST.DBDLIB’
//STEP1.CTL DD *
SD144P DFSHDC40 001000 004 08192 001 007 03000 HDAM
SD148P DFSHDC40 001500 002 04096
//TUNE2 EXEC FABTIMS,
// CYL=’10,10’, SPACE FOR WORK FILES
// PARM2=’SIZE=(MAX)’, PARM FOR DFSORT
// KEYSIN=’HPS.TEST.KEYSIN’, INPUT KEYS
// DBDLIB=’HPS.TEST.DBDLIB’
//STEP1.CTL DD *
SD144P DFSHDC20 001000 004 08192 001 007 03000 HDAM
SD148P DFSHDC20 001500 002 04096
/*
//HD Tuning Aid is being run twice for each database. Because module FABTROOT only makes one pass through the KEYSIN data set, HD Tuning Aid must be run twice. Each execution of HD Tuning Aid processes both databases once.
Because SD144P is an HIDAM DBD, it is required that all fields on the control statement be used, even if the actual DBD value is not being changed.