Citizen Bob wrote
(Citizen Bob) wrote

After getting corrupt NTFS volumes a couple times a day, including
a BSOD, I have gone a couple days with nothing going wrong. That's
because I used Perfect Disk to defrag both online and offline. I have
noticed this before. I expect in about 3-5 days I will begin seeing
the lone corruption followed in intensity by more and finally a BSOD
or two before I defrag again and it all stops for about a week.

What does that tell you about the cause of this?

I have gone several days without any corruption.

No new info, you've had that before.

I have cycled all three cloned disks. This is why
I do not believe it's the removable drive bays.

Thats a stupid way to test that possibility.

Here's what I did:

1) I changed the pagefile to 512MB/1500MB per Kony's recommendation.

It wont be that, its already FAR bigger than you
need and wont be what is corrupting the MFT.

2) I ran PerfectDisk defragment both online and offline.
In the past this has extended the period with no corruption.

You have previously claimed that you can get corruption
of the MFT just by running it. You cant have it both ways.

3) I have run CHKDSK without the F several times each day.

Presumably you mean that it doesnt report any problems.
Then you have confirmed that the corruption is occuring
at boot or shutdown time, except for the corruption you
claimed you produced by just running PD.

4) I changed the signature on each clone with Win98SE fdisk/mbr.
I also deleted all the keys in HKLM/System/MountedDevices.
Then I Installed the "new device" when Win2K was rebooted.

That's it. Nothing else is different. I have
run the same applications I always run.

We already know that the fault is intermittent.

This does not mean that corruption won't occur sometime soon.
But for now there is absolutely no corruption whatsoever.

Since you have gone that long without corruption in the past, that proves nothing.

You have to test intermittent faults more rigorously
than that, because they are intermittent.

You've now wasted FAR more time than it would have taken to
do the two basic tests that would at least distinguish between a
hardware problem and a ****ed OS install, running with the drive
directly connected and trying a clean install of XP on a spare
hard drive using the files and settings transfer wizard.

If you find that say the XP install works fine, you dont necessarily
have to use it, you can decide that you have the proof that its a
****ed OS install and put more time into a clean 2K install if you want.