After going through bug #57556, I decided to lodge my report.

Hardware: 
Two IBM SCSI Ultrastar LVD disks in ch A (id 1 and id 2)
AIC 7896 - Adaptec Bios v2.20S1B1
512 MB RAM
Intel 440LX
Dual i686 Processors (800 MHz)

Disk 1 was bootable and this is the one giving problem now. 

According to #57556 the  bug is fixed in the current release (RH9). 
In my system the bug has appeared six months after I upgraded to RH9 
from 6.2.

The boot gives 'start unit request failed' for disk 1, goes on to 
disk 2 (which is not bootable) and then hangs.

---

SCSI utility tool gave the diagnosis:

Target status = 02h (device not ready)
sense key: 02h
code: 40h
Qualifier: 85h

---

There are a few observations:

1. A week ago, a similar problem appeared but that time it was disk 
#2 that gave problem; Looking into dmesg it showed 'trying to spin up 
the drive and failed'. Fortunately disk 1 was bootable then. Now it 
is disk 1 giving the problem and the system recognizes disk 2! Hence, 
I believe the problem is not exactly a hardware (e.g. burnt disk).

2. Removing disk 2 did not help.

3. IBM's website says 'clear volatile memory' by turning it off for a 
few minutes. That didn't help.

4. One could hear the system trying to spin up the drive, through the 
clicking sound.

5. PhoenixBios default settings are used to boot. Fiddling around 
with some variables (like turning off start request) didn't help.

---

Barani