Tool Mentor: Executing a Performance Test Schedule
Purpose
This tool mentor describes how to run a schedule for performance testing using Rational LoadTest.
Related Rational Unified Process activities:
Also see the following tool mentors for additional related information:
- See Tool Mentor: Monitoring a Performance Test Schedule
- See Tool Mentor: Reporting on a Performance Test
Overview
When doing performance testing, you create and run schedules with LoadTest. A
schedule shows a hierarchical representation of the load that you want to run to
test the performance of your system. It shows such items as the user groups, the
number of users in each user group, which scripts the user groups run, and how
many times each script runs.
Using a schedule, you can:
- Run virtual user scripts and GUI scripts.
- Group scripts to emulate the actions of different types of users.
- Set the order in which scripts run.
- Synchronize users.
For information on designing and creating schedules, see the tool mentor: Creating a Schedule to Represent a Performance Test Workload.
After you have created a schedule, but before you run it, you may want to set some runtime and termination options.
- Controlling runtime information of a schedule.
- Controlling how a schedule terminates.
- Running the schedule.
1. Controlling runtime information of a
schedule. 
LoadTest lets you control the way a schedule runs. You can change the following types of runtime settings:
- How users are started ù either all at once or in groups.
- The criteria for whether a schedule passes.
- The order in which the user groups run.
- Timing information such as how long the run should be, how long to take to initialize the system, whether to suppress timing delays, and whether to initialize timestamps for each script.
The procedure for setting these
runtime options is found in Chapter 7 of the Using Rational LoadTest
manual, which can be found on the documentation CD.
2. Controlling how a schedule terminates.
LoadTest lets you set the conditions that force a schedule to terminate. For
example, you may want to stop a schedule if you discover that a large number of
users are completing abnormally, indicating that something is wrong with the
run.
The procedure for controlling how a
schedule terminates is found in Chapter 7 of the Using Rational LoadTest
manual, which can be found on the documentation CD.
3. Running the schedule.
When you run a schedule, each user executes its assigned items. The results
of running the schedule are stored in logs. After you run the schedule, you can
run reports to analyze the data stored in the logs and display the results in
the form of graphs and charts.
After you start a schedule run, LoadTest displays a Preparing to Run window.
LoadTest checks the schedule, and compiles any new or modified scripts. You can
then monitor the schedule as it runs or leave it unattended.
The procedure for running a
schedule is found in Chapter 7 of the Using Rational LoadTest manual,
which can be found on the documentation CD.
