T22: hd and/or mobo problem?

[l.butler]

Hi, let me apologize for the length of this post. I have spent the last month
troubleshooting this problem.

The problem began in early January: I (stupidly) put the machine in hibernate
mode, put it in a travel bag and went for a 7-10 minute walk to the library.
After about 5 minutes of work, I tried to save a file and got an error message
to the effect that hd was unavailable. I immediately powered off, and tried to
reboot. The typical "OS not installed, HD unavailable" message came up, so I
powered off again. Because it was very hot, I let the machine stand
for an hour, and it booted up normally.

It booted up normally for the next week, until I tried to boot up at LAX while
waiting for a flight. Same error message as above, except machine was cold.

Back home, I got the same problem. However, it booted from my Knoppix rescue CD
on first try, and then it booted from the hd afterward. As before, the problem
disappeared. In the next couple weeks, it was running for periods of 1.5, 3.5,
6.5 hrs at a stretch. I installed gkrellm to monitor temps, etc., and I had the
temp.s up to 70-72C. The fan powers on reliably, cools it down to ~45C.

The problem re-appeared a few days ago while working from home. After about 3.5
hrs of work, with temp.s consistently ~45-52C, a sound, not unlike marbles
rattling in a bag, started. The hd monitor, which, with low workloads, usually
shows a 16-25k spike every 3-5sec, showed no read-write activity on the hd.
However, the green hd light showed that the hd was being read-written as usual.

Diagnostics: I have run the full suite of PC-doctor tests (thanks, Ray) many
times without any "failures". After the last problem, I ran the hd suite again.
Although it took somewhat longer, there were no "failures".

I have gone through the hmm and followed its instructions (p. 52): 1. the OS is
installed and functioning; 2. there don't appear to be any BIOS errors; 3. hd
has been removed and reseated; 4. I have not reinstalled the OS, because this is
clearly not the problem (?).

To date, no data has been lost on the hd, except for one file that appears to
have been lost when a read-write problem happened.

Any opinions on the nature of the problem? how to fix it?

t22 stats: PIII, 40G HD, 386M RAM, CD-RW/DVD-R,
HD-installed Knoppix (~Debian/GNU linux with 2.6.18-i686 kernel)

[Stargate199]

I hate to say this about your HD buy "He's dead, Jim." This is typical of a hard drive going bad. Based on all you have said, it seems like all the mechanics of the hard drive are working fine, but the circuit boards seem to be going bad. This would account why the computer was not recognizing the HD. Luckily for you, it is still somewhat working. I would go out and buy a new hard drive (pretty cheap today, about $1 per GB). Then located a 2.5 USB hard disk enclosure. After you install Windows on your hard drive (or whatever OS of your choice). Put your old hard drive into the enclosure and connect it to your USB port. Then as quickly as you can (before the HD give out again.) transfer all of your data to the new drive.

IF there was something mechanically wrong with the drive, you would get a boot error upon OS boot. For Windows XP/2000, that would be "Cannot Locate C:\%systemroot%\system32\ntoskrnl.exe" I have seen that error on one of my desktop drives. It will not install or boot Windows, but somehow it will boot Linux. Hope some of this helps.