サポートと今すぐチャット
サポートとのチャット

Foglight 5.9.1 - Performance Tuning Field Guide

Overview Hardware and Operating System Tuning Management Server Tuning Java Virtual Machine Tuning Backend Database Tuning High Availability (HA) Tuning Agent Tuning Appendix: Performance Health Check Appendix: Analyzing a Support Bundle

Oracle Tuning

When tuning an Oracle® backend database, you should first read the information below. You want the information you obtain to be sufficient enough for you to make decisions about the initial setup and configuration of the Oracle backend. Once the backend is operational, a DBA should use Oracle tools to check the parameters of execution and adjust them based on the observed behavior.

How the Management Server Uses the Database

The following are the Foglight™ Management Server backend database usage characteristics:

•
For the most part, the Management Server performs online transaction processing (OLTP) combined with occasional small-to-medium batch operations.
•
The Management Server does not use Database Management System (DBMS) jobs.
•
The Management Server configuration tables are mainly infrequent reads/updates (tablespace group A). For more information, see Oracle Tablespaces.
•
The Management Server alarm tables are frequent inserts (alarm_*, tablespace group A). For more information, see Oracle Tablespaces.
•
The Management Server topology tables are less frequent inserts/updates (topology_*, tablespace group A). For more information, see Oracle Tablespaces.

The following sections elaborate on specific aspects of Oracle® parameters.

Index Management

We recommend you consult Oracle® analysis tools to check the effectiveness of indices used by the Foglight™ Management Server. There are no known backend operations through the Management Server that would invalidate specific indices on particular events. Indices should only be rebuilt at the discretion of the DBA, after a thorough analysis of the behavior of the backend at runtime has been performed.

 
Table 2. Oracle Database Analysis
Motivation

Oracle has the ability to analyze database tables and change its data access/management strategies.

•
A fresh, 10G or later, installation of Oracle analyzes tables automatically. The database keeps track of the tables that are being modified and analyzes heavily-updated tables more frequently. This frequency has yet to be confirmed. It may be only once per day.
•
If DBAs disable this functionality and use manually scheduled scripts, then Foglight tables definitely have to be analyzed, especially the large ones like topology, observations, and alarms.
Symptom

It seems that starting with a fresh Management Server installation and attaching a large number of agents almost immediately leads to poor database performance, which persists until statistics have been calculated.

Diagnosis / Verification

The DBA should use Oracle analysis tools to check for strategy effectiveness.

Tuning

We recommend you run database analysis periodically, if it is not happening automatically (during off hours).

Calls:

exec dbms_utility.analyze_schema(‘FOGLIGHT’, ‘COMPUTE’);

or

exec dbms_utility.analyze_schema(‘FOGLIGHT’, ‘ESTIMATE’); //

 

Note: Estimate is likely to be quicker than Compute, but it is not as thorough.

Memory

At the moment, there is no direct correlation between the amount of memory required by the Foglight™ Management Server and the amount of memory required for the database. The management tools that ship with Oracle® can provide some tuning information to help you understand how much memory the database requires at load time.

Typically, we would expect that for a Management Server heap of 1 GB, the database System Global Area (SGA) should be between 1 and 2 GB. This should be set up at the DBA’s discretion.

関連ドキュメント

The document was helpful.

評価を選択

I easily found the information I needed.

評価を選択