My friend Elaine, a systems administrator, installs servers for Windows 2000 and Windows NT 4.0 on a FAT partition. She insists that installing servers on a FAT partition gives her more freedom and better control for startup troubleshooting. After she's sure that the system is running correctly, she uses convert.exe to convert the file system to NTFS. To see if your knowledge about file-system conversions matches Elaine's, take the following test. (Answers appear on page 30.)

Questions

  1. The biggest difference between convert.exe for Win2K and NT 4.0 is:
    1. Win2K uses FAT free space for temporary files, and NT 4.0 converts files directly over FAT files.
    2. Win2K converts to NTFS from FAT16 and FAT32, and NT 4.0 converts to NTFS only from FAT16.
    3. Win2K includes a parameter that protects bootsect.dos so that Windows 9x partitions can continue to read all system files. NT 4.0 doesn't provide a similar feature.
    4. All of the above.
    5. None of the above.
      
  2. Which of the following statements about postconversion permissions is correct?
    1. After converting from FAT to NTFS, NT 4.0 changes all folder and file permissions to Everyone-Full Control.
    2. After converting from FAT to NTFS, Win2K changes all folder and file permissions to Everyone-Full Control.
    3. Both of the above.
    4. Neither of the above.
      
  3. Which of the following statements about upgrading from NT 4.0 to Win2K is correct?
    1. If you upgrade from NT 4.0 running on NTFS, Win2K automatically updates the file system to NTFS 5. 0 (NTFS5). You don't have the option of installing Win2K with FAT16 or FAT32.
    2. If you upgrade from NT 4.0 running on FAT, you can use FAT16, FAT32, NTFS4, or NTFS5 to upgrade to Win2K.
    3. Both of the above.
    4. Neither of the above.

MAY WINNERS
Congratulations to Paul Hickey of Boston and Lieven Dhaenens of Ghent, Belgium. Paul won first prize of $100 for the best solution to the May Reader Challenge. Lieven won second prize of a copy of Troubleshooting NT Logons (O'Reilly & Associates Publishing, 2000).

Problem
Peggy, a Help desk worker, received a call from a user who complained that a certain application ran every time the user booted her Windows NT computer. To solve the user's problem, Peggy checked for the application in the Startup folder in the Programs menu, but the folder was empty. She then checked the HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Run Registry key, but it didn't list the persistent program. Next, Peggy checked HKEY_CURRENT_USER\SOFTWARE\Microsoft\Windows\CurrentVersion\Run, but it also didn't list the program she was looking for. She searched the entire Registry for the keyword run and found nine Registry keys, none of which listed the persistent program.

Peggy gave up and told the user to close the program after it opened. Privately, Peggy thought that the user must have been clicking a desktop shortcut that caused the pesky application to run at startup. However, Peggy failed to check an obvious place that might list the program. What place did Peggy overlook in her search?

Solution
Hundreds of readers offered excellent answers to this problem, and almost all the answers (e.g., autoexec.bat, the Startup folder, logon scripts) were correct. The majority of readers included the most usually overlooked place to find a Load command—win.ini. My private poll of users and administrators who have faced this problem lists win.ini as the most common culprit.