I have a legacy system that runs OS/2 1.3, which dates back to 1991 and uses Cylinder-Head-Sector (CHS) disk addressing. Over time, we have migrated to IDE SSD drives and using only the first 500 Mb by restricting CHS settings to 1022 Cylinders, 14 heads, 62 sectors. This is worked until recently when we had to migrate from the original CPU Board (A) to a different CPU board (B) with a different BIOS, dating to 1998.
With CPU Board A, we can partition, format and do a Ghost restore and everything works. Using this IDE SSD, we can install CPU Board B and it properly reads the hard disk.
But when creating a new disk using CPU Board B, a partition, format and Ghost restore results in a corrupted disk image. Examining the partition table with Norton shows that the Cylinder count differs from a disk created by CPU Board A by 1 less cylinder, although the user parameters were matched in BIOS. Changing the value in BIOS by 1 still results in a corrupted image.
It's as though with CPU Board B, it's telling BIOS one thing, while MSDOS or Ghost see a different value for CHS.
Any ideas about what could be going on?
[ Updating our computer completely is an option, but costs over $1 million per machine because of all the peripherals that must also be updated. ]