[rescue] Another Reason Windows Sucks

[Sheldon T. Hall]

On Tue, 9 Jul 2002 08:34:46 +0000 (GMT) Kris Kirby <kris at catonic.net> said
...

> On Mon, 8 Jul 2002, James Sharp wrote:
>> I thought all the whiz-bang IDE drives in the laptops these days did that
>> automagically.  Once you started getting bad sectors reported to the OS,
>> then it was time to throw the disk out.
>
> There's a thread on freebsd-hackers right now that went on detailing that
> a new technique for cramming more bits onto the platter is reading the
> whole track and writing it back and changing the relavent parts. This
> means that there is no fixed sectoring system and the new track lies
> `approximately' on top of the old one.
>
> Guess what happens if you loose power during a disk write? You literally
> loose sectors that cannot be recovered because ATA got rid of the FORMAT
> op-code. The bad part is that some of the new SCSI drives are implementing
> the `feature'.

Thanks for that little detail. Have you any idea when they started
incorporating that "feature" in laptop IDE drives?  That sounds very much
like what recently happened to the drive in my company-issued FPOS Wintel
laptop.  The cooling fan crapped out, the CPU overheated, the thermo-sensor
tripped, and it turned itself off.  Instantly. Several times.  Finally, it
scrambled the FAT table, and that was, as they say, that.

I was able to re-format the drive, though, so either (a) it wasn't as hosed
as those in your example or (b) this drive (dated "jan 00") doesn't use that
technique.  Or (c) I was just lucky, I suppose.

Now, in a laptop, sudden power failures should be pretty rare, but if the
disk manufacturers are starting to use this in "real" drives, I think I
might invest in some UPS manufacturer shares....

-Shel
--
.sig: no such file