4. Operation › 4.3 Operations Reference › 4.3.3 Operational Processes, Jobs, and Steps › 4.3.3.2 Operational Jobs › 4.3.3.2.17 SPLITSMF Operational Job
4.3.3.2.17 SPLITSMF Operational Job
+-----------------------------------------------------------+
| |
| SMF Input Data Source |
| CA SMF Director Data Source |
| |
| | |
| | |
| V |
| Unit +---------+ ccc Product |
| Checkpoint-------->| SPLITSMF|<--------->Incremental |
| File +---------+ Update |
| | Checkpoint |
| | files |
| | |
| +--------------------------------------------+ |
| | | | | |
| V V V V |
| ccc product ccc product ccc product ccc product |
| IUSPLTDS IUSPLTDS IUSPLTDS IUSPLTDS |
| data file data file data file data file |
| |
+-----------------------------------------------------------+
If incremental update is enabled for one or more products
that take their input data from the common SMF data files,
then you may optionally run the SPLITSMF job one or more
times each day to prepare input data for the INCRccc jobs.
The SPLITSMF job is enabled based on the products active in
the unit database, the prefix.MICS.PARMS(JCLDEF) SMFRECORDING
option, and the individual prefix.MICS.PARMS(cccOPS)
INCRSPLIT USE parameter specifications.
The SPLITSMF job JCL is in prefix.MICS.CNTL(SPLITSMF). The
SPLITSMF job is tailored to the prefix.MICS.PARMS(INPUTRDR)
member for the unit database.
The SPLITSMF job operates as follows:
o Determine applicable products by first selecting the
products that input the common SMF input data (just like
the DAILY job DAYSMF step) and then trim this list to just
those products with INCRUPDATE YES and INCRSPLIT USE
specified in the product's prefix.MICS.PARMS(cccOPS)
parameter.
o For each applicable product, allocate the product's
incremental update checkpoint file and verify that
previous INCRccc or DAYnnn processing completed
successfully.
o Construct a composite, temporary checkpoint file by
combining the contents of the unit checkpoint file and
each applicable product's incremental update checkpoint
file. This composite checkpoint file reflects the input
data processed by prior INCRccc jobs executed today, and
is used to drop any previously processed input data from
the input SMF data files.
o For each applicable product, allocate the
prefix.MICS.ccc.IUSPLTDS data set to contain the product's
subset of the SMF input data. Data set allocation
parameters for these data sets are specified on the
INCRSPLIT USE parameter. Data set allocation parameters
default to UNIT=&DASDUNIT (see JCLDEF DASDUNIT parameter)
and SPACE=(TRK,(700,100)).
The ccc.IUSPLTDS data sets are normally allocated
DISP=(NEW,CATLG). However, if a product's ccc.IUSPLTDS
already exists, the existing data set will be allocated
DISP=MOD and the new input data will be appended to any
existing data. This is done to avoid losing data when
SPLITSMF is run prior to completing all associated INCRccc
job executions. Any duplicate input data is dropped by
standard CA MICS checkpoint and SORT NODUPS processing.
A warning message will be written to the MICSLOG when the
prefix.MICS.ccc.IUSPLTDS data set already exists.
Note: The USER.IUSPLTDS data set name and data set
allocation parameters may be overridden by defining
the &IUSUSERD and/or &IUSUSERA SAS global variables
with the desired values. The following is a sample
prefix.MICS.USER.SOURCE(#BASEXIT) specification
excerpt.
%LET IUSUSERD = MYPREFIX.MYUSER.IUSPLTDS
%LET IUSUSERA = STORCLAS=WORKDATA SPACE=(CYL,(20,5)
o Read the INPUTSMF data sets and split-out the input data
to the appropriate prefix.MICS.ccc.IUSPLTDS data sets.
This is the same basic logic as it appears in the DAILY
job's DAYSMF step.
DATABASE UPDATES
The SPLITSMF job does NOT update the unit or incremental
update checkpoint or database files. SPLITSMF simply creates
the input data files for INCRccc jobs.
DEPENDENCIES
If the SPLITSMF job malfunctions, the INCRccc jobs that input
the data sets created by the SPLITSMF job may be unable to
execute. The SPLITSMF job must successfully complete before
associated INCRccc jobs can start.
The SPLITSMF job will not be able to execute if any of the
latest DAILY, WEEKLY, MONTHLY, YEARLY, BACKUP or RESTORE
jobs, or a prior INCRccc job for any product marked for
INCRSPLIT USE, have completed unsuccessfully.
If you need to force SPLITSMF job execution to recreate a
missing prefix.MICS.ccc.IUSPLTDS data set in order to recover
from an INCRccc job failure, specify,
SYSPARM=FORCE
on the SPLITSMF step EXEC statement. SYSPARM=FORCE overrides
the abort for INCRccc job failures and allows SPLITSMF job
processing to continue.
Note: The SPLITSMF job is independent of any INCRccc jobs
for products marked as INCRSPLIT IGNORE (the default)
in prefix.MICS.PARMS(cccOPS). INCRccc job failures
for these products have NO impact on SPLITSMF job
execution.
If the SPLITSMF job is accidentally submitted for execution
twice in the same day with the same data in the input files,
the second set of INCRccc jobs may abend (U300) because the
duplicate data check mechanisms in CA MICS recognize that the
same data is being submitted again.
