Siebel Enterprise Integration Manager Administration Guide > Importing Data > Importing Legacy Data >

Importing Large Databases

Before importing a large database, such as a legacy database, you should thoroughly test your import processes. Once the test batches are loaded correctly and any data discrepancies that may affect other batches are resolved, you may want to consider importing large batches for the remaining data. Before doing so, first make sure that the Siebel database is capable of storing the volume of data, and that your resources are adequate to support the processing.

Memory Resources Needed for EIM

To achieve and maintain high performance, the database memory area needs to be large enough to hold most of the frequently accessed data in the cache. Because a very large EIM batch may flush all the data from the cache and cause performance degradation, limit EIM batch sizes so the most frequently accessed data can remain in memory.

Database Resources Needed for EIM

EIM uses database server space for the EIM tables, target base tables, secondary tables, and work areas. To make sure that an import process runs smoothly to completion, you must anticipate and plan for these space requirements. Actual requirements vary based on the RDBMS you are using and the size of the database you are populating. Work with your Siebel representative and database administrator to develop a database blueprint that addresses the following resource requirements:

• Base tables and indexes. When establishing appropriate sizes for the Siebel base tables and indexes, consider not only current size, but also reasonable growth. You should plan for future changes that may affect the database, such as organization expansion, new product lines, and company acquisitions. For more information on table sizing, see the documentation for your RDBMS.
• Secondary tables. You may be importing data from a single EIM table into multiple destination tables. For each EIM table (except EIM_note), there is a primary, or target, Siebel base table. In addition, there may be one or more secondary tables associated with the target table. Data from the EIM table may ultimately reside in one of these secondary tables.
• Database manager transaction logging area. The database manager uses a disk area to log its transactions. If you fail to set an adequately sized logging area for this operation, the database manager halts when the area runs out of space.
• Transaction rollback areas. Database resources are temporarily allocated to store intermediate results used to recover the original database state if a transaction is rolled back or aborted. Each RDBMS may use a different implementation. The amount of data processed in a transaction determines the amount of database resources required for rollback areas. Make sure that you allocate sufficient resources, or use smaller batch sizes, to handle the rollback requirements. Your database administrator can configure your database to allocate adequate transaction rollback areas.

After working with small batches to make sure that your import processes run smoothly, you may want to initiate an unattended session in which EIM runs multiple import processes to load a large database.