On Wed, 06 Apr 2011 19:53:54 -0700, Darren New wrote:

> On 4/6/2011 19:20, Jim Henderson wrote:
>> they consider a single reported *recoverable* bit error on a drive in
>> their cabinets as reason enough to replace the drive.
> 
> Google's whitepapaer basically said "Nothing says when a drive is going
> to die, but you have a +30% chance of it dying in the next six months if
> you get any read errors at all, recoverable or not."
> 
> In particular, all that S.M.A.R.T. stuff was stupid.

I'd believe that - which is why for critical data, I'd prefer an 
arrangement like the one my previous employer had with EMC to deal with 
it.

Jim

Post a reply to this message