<div class="gmail_quote">On Mon, Apr 26, 2010 at 7:38 AM, <span dir="ltr"><<a href="mailto:criggie@criggie.dyndns.org">criggie@criggie.dyndns.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div class="im">Aaron Whitehouse wrote:<br>
> It was interesting to me that SMART didn't show anything.<br>
<br>
</div>Not really - smart is like an airbag... good to have but not going to<br>
save you in itself.<br>
<br>
I have a vague memory that google did an analysis of their hard drives<br>
and found that smart only predicted failure in under 1/3 of the cases.<br></blockquote><div><br>That paper, for anyone interested: <a href="http://static.googleusercontent.com/external_content/untrusted_dlcp/labs.google.com/en//papers/disk_failures.pdf">http://static.googleusercontent.com/external_content/untrusted_dlcp/labs.google.com/en//papers/disk_failures.pdf</a><br>
<br>On SMARTs use as a predictive tool it says:<br>"Out of all failed drives, over 56% of them have no count in any of the four strong SMART signals, namely scan errors, reallocation count, offline reallocation, and probational count. In other words, models based only on those signals can never predict more than half of the failed drives. ... even when we add all remaining SMART parameters (except temperature)<br>
we still find that over 36% of all failed drives had zero counts on all variables."<br><br>Scan errors are the best correlated SMART parameter to drive failure: "after their first scan error, drives are 39 times more likely to fail within 60 days than drives with no such errors". Yet 70% of drives still survived 8 months after their first scan error.<br>
<br>Cheers,<br>Steve<br> <br></div></div>