JetStress 7795 has the first two issues fixed. http://www.microsoft.com/downloads/details.aspx?familyid=94b9810b-670e-433a-b5ef-b47054595e9c 

1. The test report may incorrectly state the test has failed in the header of the test report --- JetStress 7795 has it fixed.

  • If your test report suggests that the test has failed, you should carefully analyze the data provided in the report to ensure that the test actually failed.

2. The test database creation doesn't go to 100% --- JetStress 7795 has it fixed.

3. The test report has an incorrect total database size. --- JetStress 7830 has it fixed.

  • It calculates the total database size when you prepare databases and/or start a test-run.
  • It re-calculates the total database size if you switch the main tab user interface control even when it is running a test-run --- This is a bug --- It should only happen when it is inactive.
  • To stay away from this bug, you'd better not to switch the main tab(s) while it is running a test-run. The fix needs to go through more testing, so it won't be available soon.

4. Unhandled Exception such as System.InvalidOperationException: Cannot read Instance: E:\Data\Mount-Point.

  • There are many times Win32 PDH library, .NET Performance library, and/or disk sub-system are not successfully initialized.
  • Sometimes, you can see logical disk instance names like "HarddiskVolume7", or "HarddiskVolume8" using the PerfMon MMCsnap-in.
  • Volume mount points on disks in a server cluster may intermittently become unavailable after you install Windows Server 2003 Service Pack 1: http://support.microsoft.com/Default.aspx?kbid=898790

  • The bug description is like this: To better handle situations where logical disks are not working properly, JetStress has to verify accessing logical disk instances before it can start any test run --- The latest build (7830) properly reports these intermittent faults of mount points (or other disk volumes).