CHECK PLUS DD statements

SORTWK data sets

SORTWK data sets are the work files that BMCSORT uses. BMCSORT uses the data sets in the CHECK phase and uses them only for CHECK INDEX level 2 processing and CHECK DATA processing.

For any job in which CHECK PLUS performs a sort, you must allocate SORTWK files in one of the following ways:

• (Recommended) Have BMCSORT dynamically allocate SORTWK data sets. For more information, see Dynamically allocating SORTWK data sets.
• Explicitly specify SORTWK DD statements in your JCL. Use this option when you want to control the allocation of your SORTWK data sets. For more information, see Allocating SORTWK data sets in your JCL.

Dynamically allocating SORTWK data sets

BMCSORT dynamically allocates SORTWK data sets if they are not already allocated explicitly in your JCL (or if it needs more SORTWK data sets than you have allocated) and you have enabled dynamic allocation in BMCSORT. We recommend that you specify SORTNUM 32. 

Allocating SORTWK data sets in your JCL

If you cannot dynamically allocate SORTWK data sets, you must specify SORTWK DD statements in your JCL. The number of SORTWK files should be evenly divisible by the number of tasks, with a minimum of two sort work files per task. Available resources might limit the number of concurrent tasks to four. Therefore, if you allocate 12 SORTWK data sets, CHECK PLUS can run up to four tasks without wasting sort space.

SORTWK size for CHECK INDEX processing

To determine the SORTWK data set size for CHECK INDEX level 2 processing, use the following formula for each table:

  Use the following formula for each index:

Add all of these values and multiply by 1.2. This is the total amount of sort work space you will need. Divide this value by the number of SORTWK files and round up to the nearest cylinder boundary.

SORTWK size for CHECK DATA processing

To determine the total SORTWK data set size for CHECK DATA processing, calculate the value, divide by the number of SORTWK files, and then round up to the nearest cylinder boundary:

When determining the number of concurrent tasks to run, CHECK PLUS checks the amount of allocated SORTWK space. In these calculations, CHECK PLUS uses only the primary allocation because the secondary allocation is not guaranteed.

Restrictions

You cannot allocate a SORTWK data set as any of the following data set types:

• VIO data set
• Tape data set
• Multi-volume data set
• Data set in an SMS storage group that specifies EXTENDED FORMAT YES

BMCSORT does not support SORTWK data sets that extend beyond 65,535 tracks on a single volume.

SYSDISC data set

CHECK PLUS always requires a SYSDISC data set. CHECK PLUS writes the distinct foreign keys in RI violation to this data set during the FIX RI phase. To override the default ddname SYSDISC, use the DISCARDDN option.

For details about specifying the format of the SYSDISC data set, see the description for the DISCARDFMT option on. For information about using the CHECK PLUS provided program (ACKESQL) to run the SQL DELETE statements in this data set, see SQL batch execution program.

SYSERR data set

CHECK PLUS always requires an SYSERR data set for CHECK DATA processing. CHECK PLUS uses this work data set for error processing in the CHECK and FIX RI phases. To override the default ddname SYSERR, use the ERRDDN option.  

SYSIN data set

CHECK PLUS always requires an SYSIN data set. This input data set contains the CHECK command. The UTILINIT phase reads, parses, and verifies the CHECK command that is provided in this data set. You must specify the attributes of this data set as fixed length (RECFM is F, FB, or FBS), and the record length must be 80 columns (LRECL=80). CHECK PLUS uses only columns 1 through 72.

SYSPRINT data set

CHECK PLUS always requires an SYSPRINT data set. This output data set contains CHECK PLUS messages.

CHECK PLUS overrides any data control block attributes that you specify in your JCL with DCB=(RECFM=VBA, LRECL=137, BLKSIZE=141).

Note the considerations if you direct SYSPRINT to a tape or disk data set:

• The DSN messages from Db2, such as those from a QUIESCE utility, are lost.
• We recommend that you do not specify the BUFNO parameter on the DD statement for this data set.

For information about the level of messages that CHECK PLUS displays and how to change the message level, see Message level (MSGLEVEL).

UTPRINT data set

CHECK PLUS always requires a UTPRINT data set if sorting is necessary. The presence of this data set tells CHECK PLUS to report sort messages. However, the actual messages for each sort process appear in separate SYSnnnnn data sets, where nnnnn is a system-assigned sequential number. You cannot specify a sequential or partitioned data set for UTPRINT. UTPRINT supports only SYSOUT data sets.

Other data sets

Because it uses BMCSORT, CHECK PLUS ignores any traditional sort of routine DD statements (such as $ORTPARM and DFSPARM) that you specify.