6. Database Tailoring and Retrofitting Techniques › 6.2 Tailoring the Database › 6.2.3 Tailoring Archive Files › 6.2.3.1 Add File to Audit Archive Tape

6.2.3.1 Add File to Audit Archive Tape

To add a file to the audit archive tape, which is generated
by the CA MICS WEEKLY or the standalone AUDIT operational
job, do the following:
 
  1.  Edit the FILE statement in
      sharedprefix.MICS.GENLIB(cccGENIN).  Change the "N" in
      that file's audit archive position to "Y".
 
      Refer to Section 6.2.3 for more information on the
      GENIN FILES statement.
 
  2.  Submit the sharedprefix.MICS.CNTL(cccCGEN) job to
      regenerate the component.
 
  3.  Edit prefix.MICS.PARMS(DBMODEL) for all database units
      in which the component is installed.
 
      On the FILE statement for the file to be added, either
      omit an AUDIT keyword, in which case the definition
      from cccGENIN is taken, or define AUDIT(YES). Audit
      archive is created from the DETAIL time-span. If no
      DETAIL time-span file cycles exist, it is created from
      the DAYS time-span. To specifically create audit
      archive from the DAYS time-span, define AUDIT(DAYS).
      Define AUDIT(NO) in units where you do not want to
      create this file.
 
      Refer to Section 2.3.4.1.2 in the PIOM for more
      information on the FILE statement in DBMODEL.
 
  4.  Submit the prefix.MICS.CNTL(CYCLEGEN) job for all
      database units in which the component is installed.
 
      This job changes the component's executable code in the
      prefix.MICS.USER.SOURCE libraries and affects only the
      database units in which the component is installed.
 
  5.  Edit prefix.MICS.PARMS(JCLGENU) for all database units
      in which the component is installed.  Specify:
 
          WEEKLY AUDIT
 
  6.  Submit the prefix.MICS.CNTL(JCLGENU) job(s).
 
      This generates the WEEKLY and/or AUDIT job JCL that
      contains DD statements for the new audit tape
      configuration.
 
  7.  Submit prefix.MICS.CNTL(GDGSGEN) for all database units
      in which the component is installed.
 
      This generates audit tape file catalog indexes for the
      new audit file, if necessary.
 
For example, assume the NPANCP file example showed the file
appearing in the following archives:
 
 WEEKLY   WEEKLY    MONTHLY
 AUDIT    HISTORY   HISTORY
--------  --------  --------
  NO        YES       YES
 
Assume the desired configuration is:
 
 WEEKLY   WEEKLY    MONTHLY
 AUDIT    HISTORY   HISTORY
--------  --------  --------
  YES       YES       YES
 
The initial definition of the NPANCP FILE statement in
sharedprefix.MICS.GENLIB(NPAGENIN) is:
 
    FILE NCP 00 1 Y Y Y Y N N Y Y NCP Activity File
                            | | |
      no WEEKLY audit-------+ | |
      WEEKLY history ---------+ |
      MONTHLY history-----------+
 
    Change this statement to:
 
    FILE NCP 00 1 Y Y Y Y N Y Y Y NCP Activity File
                            |
                            +- WEEKLY audit produced
 
The number of cycles of the file that CA MICS will use as
input to the weekly audit archive is fixed at 10.  This means
the last 10 cycles of DETAIL (or DAYS if DETAIL does not
normally exist) will be copied to the new generation of the
audit archive tape when the audit is generated in the WEEKLY
or the standalone AUDIT operational job.  If both DETAIL and
DAYS are active, the DETAIL time-span files will be used.
You can specifically request audit archive from the DAYS
time-span with the AUDIT(DAYS) keyword on the FILE statement
in prefix.MICS.PARMS(DBMODEL).
 
The test of what observations to put on the audit file is:
any observation from a week other than the current week will
be copied to the audit file.  Any observation that belongs to
the current week will be ignored, and will not be copied to
the audit file.  If less than ten cycles of the file exist in
the DETAIL time-span, a data loss may occur.
 
The intent of using ten cycles of DETAIL data is to be able
to copy the entire previous week of DETAIL data to the audit
file, if the WEEKLY (or AUDIT) job was run as late as
WEDNESDAY of the following week.  The normal WEEKLY (or
AUDIT) job is run by Tuesday.  Normal cycle distribution
looks like:

 
                     Cycle number of associated data
                     as of the time the WEEKLY is run
                     on Tuesday
Day of     Week of   -+------------------------------
Week       Year       |
---------  --------  -+-
Friday     week n-2   11
Saturday              10 --+ --------+
------------------------   |         |
Sunday     week n-1   09   | copied  |
Monday          |     08   | to      |
Tuesday         |     07   | audit   | considered
Wednesday       |     06   |         | for
Thursday        |     05   |         | audit
Friday          |     04   |         |
Saturday        V     03 --+         |
------------------------             |
Sunday     week n     02             |
Monday          |     01 ------------+
Tuesday         V
 
 
Data loss can occur from such conditions as the following:
 
    - less than 9 cycles of DETAIL data are maintained
 
    - the WEEKLY (or AUDIT) job is run as late as Thursday
 
    - the CA MICS data element WEEK has been modified to have
      a value incompatible with the above structure.
 
    - the standalone AUDIT job is executed more than once in
      a given week, there is less then 10 cycles, and
      prefix.MICS.PARMS(EXECDEF) does not contain
 
           USERDEF AUDITCWK YES