Page fault when booting from mirror boot floppy

Archived from groups: microsoft.public.windowsnt.misc (More info?)

Just to clarify something in my original post...The crash screen is not a
typcial NT "blue screen of death". It is a black screen that is partially
overwritten with what looks to be some CPU register data. The screen is not
completely overwritten, as there are remnants of the "last known/good
configuration screen" still visible.

"Joseph Huber" wrote:

> I am having trouble with the mirror boot floppy test on my old DEC Venturis
> server running Windows NT 4.0 SP6a. The boot sequence crashes with a page
> fault before the expected blue screen shows up.
>
> Equipment:
> * Onboard NEC/Symbios Logic C810 SCSI controller
> - ScsiPort 1 when Windows has finished loading
> - This controller BIOS shows up first during the power-up sequence
> - (2) CD-ROM drives attached
> * Adaptec 2940U2W PCI controller
> - ScsiPort 0 when Windows has finished loading
> - This controller BIOS shows up after the NEC during the power-up sequence
> - Seagate ST19171N ID=0 (boot disk)
> - Seagate ST373307LW ID=1 (data disk)
> - Seagate ST336706LW ID=4 (mirror disk)
> - Seagate ST318436LW ID=9 (data disk)
> - HP C5713A ID=15 (tape drive)
>
> The computer normally boots as "multi(0)", but since my mirror is on the
> third disk, I believe that I cannot use "multi(0)" on the boot floppy and
> instead have to use "scsi(x)".
>
> The ARC path in boot.ini on my floppy is:
> scsi(0)disk(4)rdisk(0)partition(1)\WINNT351="Windows NT Server Version 4.00
> Mirror "
>
> The ARC path in boot.ini on ST19171N hard drive is:
> multi(0)disk(0)rdisk(0)partition(1)\WINNT351="Windows NT Server Version 4.00"
>
> I have a copy of Aic78u2.sys on the boot disk as ntbootdd.sys. There is
> nothing on the mirror disk except the mirror partition, and Disk
> Administrator indicates that the mirror is healthy.
>
> When booting, the screen goes black, and the (4) disk drive LEDs blink a
> little bit. It appears that the "last known configuration" screen shows up
> for just an instant, and is then overwritten by a page fault screen, with a
> bunch of register data.
>
> Any ideas?? Thanks in advance for any help!!
>