Journaling is a foundational function within the IBM i operating system and is generally used for recovery purposes. If a failure occurs during a database operation a properly configured system should not lose any transactions. Journaling is also required for commitment control as journal entries are used if transactions must be rolled back. In addition, you can also journal objects other than files, such as libraries, data queues, and data areas. A special system-supplied audit journal (QAUDJRN) can also be used to track a large number of system activities that may be required for auditing purposes. Finally, journals are commonly used for software-based replication. With journaling being the foundation for so many functions, it is essential that your journal environment be properly configured to ensure optimal performance.
In addition to journal features, IBM i also provides a very efficient performance data collector called Collection Services. Collection Services collects performance metrics about a wide variety of activities on the system, from hardware resource utilization (CPU, memory, and disk), as well as an extensive set of system metrics, including journal metrics. Every job that runs is recorded by Collection Services. The Performance Data Investigator in IBM Navigator for i provides many charts where you can visualize Collection Services data. Using the Performance Data Investigator can help you make fact-based decisions for performance tuning and optimization.
This whitepaper reviews key configuration parameters closely associated with journaling and provides information on why you may want to select various settings. While this paper does not go into the details of setting up journaling, it does discuss key parameters as they pertain to performance. The primary focus is on remote journals and performance with software replication, although many of these considerations apply to local journals as well.