Previous Topic: 4.3.13.1 The CA MICS Audit Archive Tape

Next Topic: 4.3.13.3 Using CA MICS Archive Tapes With MICF

4.3.13.2 The CA MICS History Archive Tapes


The history archive tapes are created during both weekly and
monthly processing.  Based on your specifications in
prefix.MICS.PARMS(JCLDEF), Archive Weekly History processing
will occur during the WEEKLY job or in the stand-alone HISTW
operational job, and Archive Monthly History processing will
occur during the MONTHLY job or in the stand-alone HISTM
operational job.  The history archive tapes contain
observations from the most current cycles of user-designated
WEEKS and MONTHS database unit files, respectively.


ESTABLISHING HISTORY ARCHIVING

History archiving is controlled by several CA MICS statements
and parameters.  The COMPJOB and FILE statements in
sharedprefix.MICS.GENLIB (cccGENIN), where "ccc" is the
three-character product identifier, determine if weekly
and/or monthly archiving files can be created.  The COMPJOB
statement has indicators that show whether the product is
eligible to create weekly and/or monthly archiving files.
And the FILE statement determines whether a product file is
eligible for weekly and/or monthly archiving.  The ADM files
are not supported for archiving.  You must specify parameters
in the JCLDEF and DBMODEL members of prefix.MICS.PARMS to
complete the establishment of history archiving.  These
parameters determine the retention limits of the physical
tape reel, the specific weekly and/or monthly files to be
archived, and the maximum number of file cycles to be stored
on tape.

The JCLDEF parameters are as follows:

o  The ARCHIVE HISTW YES parameter specifies that weekly
   history archiving is to take place.

   -  If you specify ARCHIVE HISTW YES JOB, weekly history
      archiving occurs in the stand-alone HISTW operational
      job.

   -  If you specify ARCHIVE HISTW YES JOB AUTOSUBMIT, the
      WEEK300 step of the WEEKLY job automatically submits
      the HISTW operational job to perform weekly history
      archiving.

   -  If you specify ARCHIVE HISTW YES STEP or take the
      default (i.e., ARCHIVE HISTW YES), weekly history
      archiving takes place in the WEEK300 step of the WEEKLY
      job.

o  The ARCHIVE HISTM YES parameter specifies that monthly
   history archiving is to take place.

   -  If you specify ARCHIVE HISTM YES JOB, monthly history
      archiving occurs in the stand-alone HISTW operational
      job.

   -  If you specify ARCHIVE HISTM YES JOB AUTOSUBMIT, the
      MONTH300 step of the MONTHLY job automatically submits
      the HISTW operational job to perform monthly history
      archiving.

   -  If you specify ARCHIVE HISTM YES STEP or take the
      default (i.e., ARCHIVE HISTM YES), monthly history
      archiving takes place in the MONTH300 step of the
      MONTHLY job.

o  The TAPEPREFIX parameter specifies the high-level
   qualifier of the data set names on the history archive
   tape.

o  The HISTWGDG parameter specifies the maximum number of
   generation data group (GDG) entries that CA MICS should
   build for the weekly history archive tape and establishes
   a default retention period.  Unless there is an overriding
   specification in the EXPDT parameter of JCLDEF, CA MICS
   will automatically multiply this number by 7 days and
   determine the retention period of the tape.  For example,
   if HISTWGDG 7 is specified in JCLDEF, CA MICS will
   generate a retention period of 49 days.  This retention
   period will apply to each file that is archived during the
   WEEK300 step.

o  The HISTMGDG parameter specifies the maximum number of
   generation data group (GDG) entries that CA MICS should
   build for the monthly history archive tape and establishes
   a default retention period.  Unless there is an overriding
   specification in the EXPDT parameter of JCLDEF, CA MICS
   will automatically multiply this number by 31 days.  For
   example, if HISTMGDG 5 is specified in JCLDEF, CA MICS
   will generate a retention period of 155 days.  This
   retention period will apply to each file that is archived
   during the MONTH300 step.

HISTWGDG and HISTMGDG do NOT equal the maximum number of
database file cycles to be stored on the tape.  The number of
cycles to be kept is determined by the values specified in
the following DBMODEL parameters:

o  The "wha" parameter on each FILE statement in
   prefix.MICS.PARMS(DBMODEL) specifies the maximum number of
   WEEKS cycles that are to be stored on the history archive
   tape. If the number is zero, the file is not archived.

o  The "mha" parameter on each FILE statement in
   prefix.MICS.PARMS(DBMODEL) specifies the maximum number of
   MONTH cycles that are to be stored on the history archive
   tape. If the number is zero, the file is not archived.

The"wha" and "mha" parameters for an active file can
easily be changed by the CA MICS System Administrator.
Simply alter the DBMODEL parameters, then submit
prefix.MICS.CNTL(CYCLEGEN).  Section 6.2.3, Tailoring
Archive Files, in the System Modification Guide explains how
to add and delete history archive files.


HOW THE HISTORY ARCHIVE TAPES ARE CREATED

The weekly and monthly history archive tapes are created in
the HISTW job or in step WEEK300 of the WEEKLY job (for
weekly history), and in the HISTM job or in step MONTH300 of
the MONTHLY job (for monthly history), respectively.  The
logic used to create the tapes is described below:

    Old master + updates = new master

The old master is the (0) generation level of a history
archive file.  It resides on tape and contains observations
from up to "n" cycles, where "n" is the limit specified in
the "wha" or "mha" parameter in DBMODEL.  The
observations in these cycles have been concatenated into one
SAS data set on tape.

The updates are the observations in the 01 cycle of the
database unit file.

The new master is the (+1) generation level of the history
archive file.  A SAS SET statement is used to concatenate the
observations from the 01 timespan cycle with those on the
old master tape.  Any observation on the old master tape that
exceeds the "wha" or "mha" limit is dropped during the
DATA step.  Note that the new history archive file is neither
summarized nor sorted during this process (see HISTORY
ARCHIVE RETRIEVAL CONSIDERATIONS below).

Each file on the weekly history archive tape is entered in
the OS catalog as:

    tapeprefix.MICS.HISTW.iiifff.GggggV00

Each file on the monthly history archive tape is entered in
the OS catalog as:

    tapeprefix.MICS.HISTM.iiifff.GggggV00

where "tapeprefix" is the high-level qualifier specified in
the TAPEPREFIX parameter of prefix.MICS.PARMS(JCLDEF),
"iiifff" is the information area and filename, and "gggg" is
the absolute number of a generation data group (GDG).


RETRIEVING DATA FROM THE HISTORY ARCHIVE TAPES

The history archive tapes may be accessed through a MICF
inquiry or a batch job.  Refer to Section 4.3.13.3, Using
CA MICS Archive Tapes With MICF, for more details about
accessing the archive tapes in MICF inquiries.

The JCL specifications that are used to access the files on
the history archive tapes are different from those used to
access the files on the database unit.  Database unit files
are referenced as:

    //S01 EXEC MICSSHRx
    //SYSIN  DD *
    DATA;
      SET &iiit..HARCPUnn;

where x is the database unit ID, "&iiit" is a CA MICS/SAS
macro variable representing the information area (iii),
timespan (t) of the file to be accessed, and nnn is cycle.
Coding the macro variable as compared to a timespan ddname
eliminates the need to know if a particular file has been
defined to DETAIL, DETAIL2 or another split database file.
Refer to Section 2.3.3.3.2.4 for information about the
CA MICS Database Split Table.

Files on the weekly history archive tape are referenced as:

    //S01 EXEC MICSNDBx
    //ddname DD DSN=tapeprefix.MICS.HISTW.iiifff<gdgnumber>,
    // DISP=OLD
    //SYSIN  DD *
    DATA;
      SET ddname.iiifff;

Files on the monthly history archive tape are referenced in a
similar manner:

    //ddname DD DSN=tapeprefix.MICS.HISTM.iiifff<gdgnumber>,
    // DISP=OLD

where x is the database unit ID, "ddname" is assigned by
you, "tapeprefix" is the high-level qualifier specified in
the TAPEPREFIX parameter, "iiifff" is the information area
and filename, and "<gdgnumber>" is either the absolute GDG
number or the relative GDG number.  Note that the timespan
and cycle number are not used to access these files.  The
observations from all of the timespan cycles have been
concatenated into one SAS data set on the tape.

Let's look at an example.  If you specified twelve (12)
generations for the HISTMGDG parameter and "PRIMARY" for the
TAPEPREFIX parameter, you could retrieve the BATJOB file from
the current monthly history archive tape by its absolute
generation number:

    //RETRIEVE DD DSN=PRIMARY.MICS.HISTM.BATJOB.G0012V00,
    // DISP=OLD

or by its relative generation number:

    //RETRIEVE DD DSN=PRIMARY.MICS.HISTM.BATJOB(0),DISP=OLD


HISTORY ARCHIVE RETRIEVAL CONSIDERATIONS

1.  You can reference only one file on a tape volume at a
    time in a SAS DATA step or procedure.  If you need more
    than one file from a history archive tape, use the SAS
    COPY procedure to copy the desired files to disk; then
    use the disk files for your requirements.

2.  Remember that each file on the history archive tape
    contains observations from multiple cycles of weekly (or
    monthly) data.  The majority of these observations
    duplicate data contained on the (-1) and older
    generations of the history archive tape because they are
    transferred from the old master to the new master until
    they reached their cutoff period.  You should only need
    to reference one generation of a history archive file
    since it would contain all information back to the cutoff
    date as defined in DBMODEL.

3.  We mentioned above that files on the history archive tape
    are not summarized as they are written to tape.  That is,
    the tape contains summarized data, but since it was built
    from multiple CA MICS file cycles, there may be multiple
    records for a given key combination (e.g., year, month,
    zone, etc.).  You should use the CA MICS Summarization
    Facility to resummarize the data before using it for your
    requirements.