Notes

Outline

Multithreaded Server Tuning Issues

Objectives After completing this lesson, you should be able to do the following: Identify issues associated with managing users in a multithreaded server environment Diagnose and resolve performance issues with multithreaded server processes Configure the multithreaded server environment to optimize performance

Overview

Multithreaded Server Characteristics Users can share processes Supports NET8i functionality Increases number of concurrent users Is most useful on: UNIX systems Other servers with remote clients Incurs some CPU overhead

Configuring the Multithreaded Server NET8i listener.ora tnsnames.ora MTS instance parameters:

Monitoring Dispatchers Identify contention for dispatchers by checking: Busy rates Dispatcher waiting time

Monitoring Dispatchers Check for dispatcher contention Dynamically add or remove dispatchers Performance Manager predefined charts: Dispatcher and Queue

Monitoring Shared Servers Oracle8i starts up shared servers dynamically. Check for shared server process contention Dynamically add or remove shared servers Use Performance Manager charts: Shared Server Queue

Monitoring Process Usage The V$CIRCUIT view: Server address Dispatcher address User session address Performance Manager charts: predefined charts Process, Circuits

Shared Servers and Memory Usage Some user information goes into the shared pool Overall memory demand should still decrease Shared servers use UGA for sorts UGA stored in large pool if configured

Possible Problems Net8i listener is not running. The MTS initialization parameters are set incorrectly. The dispatcher process has been terminated. The DBA does not have a dedicated connection. The PROCESSES parameter is too low.

Obtaining Dictionary Information

Summary In this lesson, you should have learned that: MTS is a resource-sharing configuration. MTS is not intended for batch processing or decision support. MTS requires a Net8i listener. You can monitor dispatcher and server usage. The Oracle8i Server manages shared servers dynamically.