There are some differences between two test run types (performance and stress):

• Tuning phase uses different disk latency thresholds. See the JetStress documentation for the details.
  (the next web release of JetStressUI will allow users to edit the default disk latency thresholds in JS_Config.xml. But, this should be used with caution since high thresholds can break down the disk sub-system).
• HTML report uses different criteria about test success/failure of disk latency, cpu, and memory performance counters.
• Run-time for stress test run is 24 hours.

IMHO,

• Performance test is a validation test to test what it does what it must do --- JetStress tries to achieve the planned Exchange I/O profile (IOPS, read/write ratios in log and database drives, etc).
• Stress test is a falsification test to test what it doesn't do what it must not do --- JetStress uses some high latency to put the system under stress and verify that it doesn't do what it must not do, e.g. bad database page checksums.
  (the next web release will sync its default latency threads w/ Exchange 2003 troubleshooting guide).

Maximum I/O throughput tuning type will be available with the next web release (see the following bitmap):

This posting is provided "AS IS" with no warranties, and confers no rights.