Originally posted by: Smilin
I used an internal version of a utility called chkreg.
In your case this may have been unncecessary. The XP & 2003 version of regedit has the ability to correct minor corruption. I didn't find any corruption when I ran chkreg after using the XP regedit to load the hive.
Just curious, do you think that MS will ever make that tool available to end-users? The reason I ask is, I recently (well, about 2.5 months ago) had a similar issue with a corrupt SYSTEM registry hive in W2K. I downloaded MS's W2K registry-repair tool, that runs from a modified set of WinXP Pro install floppies.
Unfortunately, that tool claimed to repair the registry, but didn't seem to actually do that much. W2K still couldn't load. I tried "repairing" both the SYSTEM and the SYSTEM.ALT hives, but no-go on either of them. I managed to get my system running again by restoring a 6-month old copy of my SYSTEM registry hive from a backup. (Backups are a Good Thing.

)
The only reason I ask is, I've used the registry-repair tool that is built into Win98se many times on client's systems to fix problems, and it has never failed to work. I was rather surprised the W2K didn't ship with such a tool, but pleased to find out that MS offered one for download. But I was again a bit surprised, that it apparently didn't work. If they have a newer version that works better, it would seem prudent to offer that version to end-users.
Do you think that it would be wise (or possible), to install the WinXP/W2K3 version of REGEDIT.EXE into a W2K installation, purely for repair purposes? Or do you need to be booted into WinXP to use it? (possible OS version dependencies?)
PS. Thanks for the detailed explaination of the failure-mode that can occur when userinit.exe fails. I learned that the hard way when Ghosting over an installed OS, and leaving the "old" disk present on the system while booting the "new" disk, as Windows has sticky drive-letter mappings, and will continue to use the "old" disk for some things. If the "new" disk was booted, even once, with the "old" disk also present, then subsequently booting with the "old" disk removed from the system, tends to result in the failure mode you described. I just never knew what
exactly was going on behind-the-scenes when that happened.