Region size considerations

The region size used in production environments is typically larger than the size shown in the examples.

For recoveries that process large amounts of data, you might need to increase the region size to avoid abnormal or early termination that might result from storage shortages.

When setting the region size parameter in the startup JCL for both MAS and RSS, use the following procedures. These procedures will determine the probable number of log records to be processed during recovery and will set a region size for the master and subordinate address spaces:

1. Check the summary report for a prior recovery of the database data sets currently being recovered.

   If a summary report is not available, you can use reports from recoveries of other databases.

   If you use such alternate reports, they should reflect recoveries with similar update patterns as the databases currently being recovered.

   You should also use reports from recoveries of other HIDAM, PHDAM, or Fast Path areas, for example, rather than using databases with organizations that are very different from the ones being recovered.

2. Add records processed for log data sets listed in the summary report.
3. Determine the average number of the processed records.
4. Run a recovery with the VERIFY function to determine the number of log data sets to be processed.
5. Multiply the average number of records found in Step 3 by the number of log data sets found in Step 4.
6. Use the following table to determine the value for the region size:

   Table 1. Guidelines for determining region size for master and subordinate address spaces

   Number of records processed	Region size
   <35 million log records	32M - 256M
   Between 35 million and 50 million log records	At least 256M
   >50 million log records	At least 512M