Re: Backing up data without errors?

In article <f716986b-b69c-45ab-b66c-53e348db6edc@xxxxxxxxxxxxxxxxxxxxxxx
s.com>, A. J. Moss <ajmoss_throwaway_account_001@xxxxxxxxxxxxx> writes

The odds of two random glitches occurring
in exactly the same place in a multi-GB file are remote.

Agreed.

Since the same errors do not crop up in the same place in
successive runs of "comp", I assume it's just the local copy
of the contents of F: (in the computer's RAM) that has been
read incorrectly, just this once.

At this point I'd be running Memtest86+ for several hours, followed by
Prime95's torture test to give the CPU and memory a workout. Have you
upgraded the memory in your laptop?

Assuming that *every* read operation I perform is slightly
imperfect (an error rate of 1 in 10^12 bits isn't too bad),
what method should I be using (e.g. the repeated running of
"comp") to ensure the data on E: is bitwise identical to the
data on F:?

You can't. The bit error rate doesn't matter: you should get an error-
free comparison every time. If you don't, there is something wrong.
What if, for instance, Windows is introducing subtle errors into files
it writes to disk? Your system will be gradually becoming more and more
corrupt.

I'd start by running memory tests as above, then proceed to unplugging
the external USB devices and experiment with comparing files on the
local disk to see if the errors are being introduced by USB.

As an aside, I didn't know about the 'comp' command, having always used
'fc' up to now. Like you, I frequently use fc to ensure files I've
written to another disk have been written correctly.

--
(\__/) Bunny says NO to Windows Vista!
(='.'=) http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.html
(")_(")

.

Loading