OSI/TS Configuration and Management Manual

Performing Monitoring and Troubleshooting Guide
OSI/TS Configuration and Management Manual—424831-001
4-3
Monitoring Performance
CPU, disk, or communications line, or by scheduling parts of the workload to run at
different times of the day.
The operating attribute values can also cause performance problems if they are set
incorrectly or are set at non-optimum values for your system. These attribute values are
set at system generation and in the configuration. In some situations, changing an
attribute may not substantially affect performance unless the congestion of the
overcommitted resource or resources is reduced.
Monitoring Performance
Once the OSI/TS subsystem is installed and operational, you should begin procedures
for ongoing monitoring. Monitoring should be a regular activity that continues for the
life of the system.
At first, monitoring should be frequent and very comprehensive. When the system
stabilizes, you can reduce monitoring activity. Your monitoring activity must always be
frequent enough and comprehensive enough to detect problems before they affect
system users. To effectively monitor performance on your system, you need to
determine the following:
•
What tools to use
•
What to monitor
•
When to monitor
What Tools to Use
You can monitor performance using various hardware and software tools, through either
operator-initiated actions or program-initiated actions. Hardware monitoring is
performed using response-time monitors, line monitors, and other hardware monitoring
tools (external to Compaq products) to determine what is happening in the CPU or on
the communications lines. Software monitoring tools include traces, SCF, DSM,
Measure, and event messages.
What Performance Indicators to Monitor
Your goal in performance analysis and monitoring is to define both the adequacy and the
limits of your system. These activities, however, must be evaluated in terms of the
performance benefit that can be obtained versus the expense of the resources.
To improve performance, it is necessary to obtain a comprehensive picture of the
subsystem, the major applications, and the resources used by each of them. This
information can be divided into four types of data:
•
Current system configuration
You can determine the current system configuration and OSI/TS subsystem
attributes using the SCF LISTDEV, SCF INFO, and SCF STATUS commands.
These commands provide information on the number of processes and other
attribute values.