Reorganization of Special Tablespaces (SAP Library - Database Administration (Oracle) with SAPDBA)

Reorganization of Special Tablespaces

When reorganizing the following tablespaces, you might run into problems:

PSAPPOOLD: This tablespace contains the ATAB table with a very large number of records (around 1 million).

The import - and thus the reorganization - of this tablespace may take up to 10 hours, depending on your operating system. The run time basically depends on the import of the table data. You can influence the duration of the import by performing the reorganization of PSAPPOOLD with a modified import script. You can find detailed information about this in the SAPNet Notes database or in the online SAPDBA help.

Furthermore, if you operate a productive database in ARCHIVELOG mode, some 1-2 GB of free space will be required in the archiving directory of the offline redo log files ( <ORACLE_HOME>/saparch ). Therefore it can be an advantage to run the database in NOARCHIVELOG mode during the reorganization of PSAPPOOLD.
See Performance Maximization during Reorganization. However, after the reorganization has been performed, a backup of the entire database is necessary.

PSAPSTABD: This tablespace usually contains the most tables with the highest total number of extents, constraints, etc.

The size of the individual Dictionary caches must be large enough to avoid ORACLE errors (for example, errors in recursive calls) during an operation on the tablespace. ORACLE itself manages the individual caches dynamically. To enable this, the init.ora parameter shared_pool_size must be defined appropriately large. No bottlenecks should normally occur.

PSAPBTABD: This tablespace grows very quickly. To avoid frequent reorganization whenever possible, we recommend regular use of the Command Option -next.

Also refer to Performance Maximization during Reorganization for information about avoiding problems during the reorganization of these tablespaces. Also use the SAPDBA online help and the SAPNet Notes database for current information about maximizing performance.