Whereas applications that require recording of System and application event data into the ENFDB, it is necessary to know how many events are created on your system. Knowing how many events are captured over a period of time will help you determine how large the ENF database should be allocated to avoid space issues. This process should only take into account events that are set to *RECORD*.
- Allocate the ENFDB temporarily at 500 Cylinders.
- Allow ENF to run with this ENFDB for several days, allowing for the ENF archive process to occur daily at a set time
See ENFPARM option BACKUP(0100). This process issues WTO CAS9373I Messages. Note this provides a count of events
since the last time the archive process has run, not a running total.
- After several days get an average count of total number of event records purged each day.
- Build a new ENFDB using the following formula:
First add the total number of events seen in the CAS9373I messages.
CAS9350I - CA-ENF Database archive commencing
CAS9373I - Purge complete for event JOBINIT, 075000 records deleted
CAS9373I - Purge complete for event JOBPURGE, 075000 records deleted
CAS9373I - Purge complete for event JOBTERM, 075000 records deleted
CAS9373I - Purge complete for event STEPTERM, 075000 records deleted
CAS9351I - CA-ENF Database archive complete, 00000012 records written
For simplicity assume that number is 300,000.
300,000 + 50% = 450,000. Allow for increase in processing at year / quarter end and allow for workspace.
450,000 * #Days = ????????. If RETPD is set to 1 day, then set #Days to 2 as it will be in the 2cnd 48hrs
that events from the 1st 24hrs will reach RETPD(1). Similarly if RETPD is set
to 2 days, then set #Days to 3 and so on.
450,000 * 2 = 900,000. We'll use 2 here since RETPD(1).
900,000 / 3500 = 257 Approximately 3500 events records per cylinder.
257 + 20 = 277 Allow extra space for DCM modules, and round up to nearest 10.
So in this case the ENFDB should be 280 cylinders.