Table of Contents

Search

  1. Preface
  2. Backing Up and Restoring the Data Vault
  3. Configuring Centera as a Remote Data Vault Store
  4. Configuring Data Archive for a Legacy Source Using Legacy Adapters
  5. Data Archive Seamless Access for PeopleSoft
  6. Data Archive Transaction Restore API
  7. Dropping and Truncating Partitions in Data Archive
  8. How to Create Business Rules to Archive and Purge Transactional Data
  9. How to Uninstall Data Archive 5.1
  10. How to Uninstall Data Archive 5.3
  11. How to Use Scripts to Change Database User Passwords in the ILM Repository
  12. IBM DB2 Database Connectivity and Setup for Data Archive
  13. Installing Data Visualization
  14. Integrating Third-Party Schedulers in ILM Engine
  15. Parallel Processing in Data Archive
  16. Seamless Access Configuration for Siebel Applications
  17. Seamless Access Setup for Oracle E-Business Suite
  18. Using the Data Vault Service JDBC Driver to Connect to the Data Vault
  19. Using Multiple Engines in an ILM Environment
  20. Using PowerExchange ODBC Connections in a Data Archive Retirement Project
  21. Discovering Foreign Key Relationships in Enterprise Data Manager

Data Archive How-To Guide

Data Archive How-To Guide

Degree of Parallelism

Degree of Parallelism

In general, the possible degree of parallelism depends on the following factors:
  • The number of CPUs available on the database server. Use the following query to find out about the number of CPUs available to an Oracle instance:
    select value from v$parameter where name = 'cpu_count';
  • The setting of the parallel_max_servers initialization parameter. Use the following query to check the current setting:
    select value from v$parameter where name = 'parallel_max_servers';
  • Set this parameter to at least twice the value of the degree of parallelism chosen. For example, to execute a delete statement with a degree of 8, Oracle needs to spawn 16 parallel threads, where 8 threads work on the necessary select to retrieve the rows and 8 perform the delete.