HP 3000 Manuals

Status and Error Tracking [ Information Access Server: Database Administration ] MPE/iX 5.0 Documentation


Information Access Server: Database Administration

Status and Error Tracking 

Error reporting is handled differently depending on how the command is
issued.  If the command is issued interactively on the HP 3000 and an
error occurs, then a message indicating its nature is displayed on the
screen.  When run as a batch stream, these messages are sent to the
standard list device ($STDLIST) that will be a spoolfile.

Status and error messages are also stored in the JOB-STATUS detail
dataset of the HDPENV database.

Here are the items in the dataset:

    USER-NAME      -- Access Server user
    JOB-NUMBER     -- MPE job number
    JOB-NAME       -- name assigned to job
    DATE-STAMP     -- date associated with messages
    TIME-STAMP     -- time associated with messages
    STATUS-MESSAGE -- status and error messages
    ACCESS-GROUP   -- user's access group
    JS-DATE        -- date in calendar format

When Access Server is installed, an IMAGE table with the name JOB-STATUS
is automatically configured as a public table.  This table allows users
to check their job status.

Alternatively, you can define view tables from this JOB-STATUS table.
These view tables can specify a user or an access group in the Where
Clause.  You can then assign table security to allow users to check only
their own status information.

Once these tables are configured, your users can access them to review
information pertaining to their batch jobs.

Clearing the JOB-STATUS Dataset 

As each batch job that is initiated from the PC is executed, status
information is recorded in the JOB-STATUS dataset of the HDPENV database.
The utility HDPUTIL is used to delete or show the records in the
JOB-STATUS dataset.  Here is the :RUN string:

:RUN HDPUTIL.PPC.SYS;

   INFO="[USER=name|*] [ON|BEFORE (mm/dd/yy)]";[PARM=2]

USER= - name of the user (* means all users)

ON or BEFORE (mm/dd/yy)--applies to user on or
before date

PARM=2 - shows records (omitting PARM=2 deletes
them)

The UDC command JSPURGE has been developed to make it easier to
periodically clear the status information and thus manage these entries
more efficiently.  The syntax for this UDC is shown below:

        :JSPURGE "USER=name|* [ON|BEFORE] date" [SHOW]

You can either delete all status entries for a particular user by
supplying the user name, or you can use the asterisk (*) instead to
delete all entries from the dataset.  For example:

             :JSPURGE "USER=*"

The ON and BEFORE options are used to purge entries on or before a
particular date:

             :JSPURGE "USER=sam ON 9/11/87"

This example will purge all entries in the JOB-STATUS dataset for user
sam on September 11, 1987.

If you just want to look at the status entries but do not want them
purged, you can use the SHOW option by itself:

             :JSPURGE "USER=sam" SHOW

Each time a purge is performed, all of the deleted information is stored
in a file called JSLOG. This file is updated and overwritten each time a
purge is invoked.  If you want to keep a detailed record of purges, you
can redirect the output to a printer or print the JSLOG file.


NOTE If you want status and error messages to be output to the JOB-STATUS table you can run HDPBATCH with PARM=1 as shown below: :RUN HDPBATCH.PPC.SYS;PARM=1 The status of the job can then be checked by examining the JOB-STATUS table.


MPE/iX 5.0 Documentation