At the recent storage conference, a Google scientist presented data demonstrating that if one drive fails in a raid array, you will get another failure in 2 to 4 weeks, I think it was, assuming they were put in at the same time. I've seen it happen, years ago on an NT4 system, where we were using the (the new) Barracudas. It's a scary thought.

>I just hope that this is just plain bad luck on one individual. So far I've been lucky enough to never lose something really valuable. But I'll keep in mind your warnings.
>
>>Several years ago (NT4 workstation) I has heading off to DevCon in Palm Springs, and then nearby to lead a training seminar. My computer crashed the day before leaving. No problem, says I: I have double backups. Both were bad. The tapes could not be recovered by a leading recovery company. I spent DevCon feverishly working on code...
>>
>>I now use a combination of raid 0 (mirroring) on my workstation, and source control on a server in a co-lo center. I've lost single drives in my mirrors, but never both at once (but often with a few months of each other). My weak point is that my notebook is not mirrored, so source control is my only backup for that (although I have a USB drive available, and really should Ghost the drive weakly, just to save the restore time).
>>
>>Hank
>>
>>><snip>
>>>
>>>>Computers have many variables. You attempt to cover all bases and something will happen outside of your control. You end up being paranoid. What if my hard disk fails? What if my backup fails? What if by back up back up fails? It is all such fun! :)
>>>
>>>Why did you do that? I never doubted my backups before reading this. Thanks for turning me also into a paranoid.
>>>
>>>My day looked so good until now.