4. Operation › 4.10 Integration with CA SMF Director › 4.10.1 CA SMF Director Prerequisites

4.10.1 CA SMF Director Prerequisites

o CA SMF Director must be managing your data center's SMF
  dump processing.

o The CA SMF Director dump process must be executed with the
  SPLIT statements as well as the INDEXDSN and INDEXDUP
  parameters identifying the primary and duplicate split
  indices respectively.

o All the split data sets created must be cataloged data
  sets.

o The CA SMF Director INDEXLIM parameter must be defined
  to set a limit on the number of entries retained in the
  index.  This number should coincide with the CA MICS
  SMFDJHLLIMIT value defined in prefix.MICS.PARMS member
  EXECDEF.  See section 2.3.5 for more information on the
  SMFDJHLLIMIT parameter and its affect on CA MICS input
  processing.

For information on the SMF record types processed by each CA
MICS product, see the "Data Sources" section of the related
component guide or the sharedprefix.MICS.PARMS(cccSMFD)
member for each CA MICS installed component.

Note: To get the most benefit from CA SMF Director and CA
MICS integration, SMF records processed by different
components should be recorded to separate CA SMF Director
duplicate split indices.  For example, if SMF and CICS reside
in the same unit, each split statement within the CA SMF
Director dump job should refer to a unique CA SMF Director
duplicate split index for recording the SMF split data sets.
This allows CA MICS to only process the data required by each
component step.  Otherwise, DAYSMF would still be needed to
further split the records to avoid each component step from
reading unnecessary SMF data sets.

Below is a sample of a CA SMF Director split statement that
demonstrates the use of a discrete duplicate split index for
SMF records processed by different components.

SPLIT ALL SELECT(70:79) TODD(RMF)
   INDEXDSN(hlq.RMF.SMFDINDX)
   TIME(0000) TIMEGDG(hlq.RMFONLY) TIMEUNIT(3390)
   TIMEVOL(SMFD01) TIMESPACE(CYL,500,200)
   INDEXDUP(INDEXRMF).


SPLIT ALL SELECT(110) TODD(CICS)
   INDEXDSN(hlq.CIC.SMFDINDX)
   TIME(0000) TIMEGDG(hlq.CIC110) TIMEUNIT(VTAPE)
   INDEXDUP(INDEXCIC).

Note: The INDEXDUP keyword is included on each SPLIT
statement and each refers to a unique DD statement that, in
turn, describes a unique duplicate split index data set.
This enables each CA MICS component step to input the
duplicate split index into the corresponding component step
without processing unneeded record types.

If, in the above example, each split data set index entry was
recorded to the same duplicate split index, the duplicate
split index should be defined to DAYSMF.  Otherwise, each
component step would read ALL the eligible input files
defined in the duplicate split index, which would result in
unnecessary processing.  The best practice would be to
separate the duplicate split index entries by component to
maximize the benefit of CA SMF Director.

For more information on implementation, see section 4.10.3.