On Mon, Oct 27, 2008 at 11:16 AM, John Drescher <span dir="ltr"><<a href="mailto:drescherjm@gmail.com">drescherjm@gmail.com</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> They also found that average utilization doesn't correlate with drive<br>
> failures unless the drive is very new (< 6 months) or very old (> 5 years).<br>
> Another interesting tidbit is that only about half of the drives that failed<br>
> had warnings from failed SMART tests.<br>
><br>
I have never had SMART warn me that a drive was going bad. And I have<br>
had 20+ disk failures out of around 200 in the last 12 years at work.<br>
<br>
</blockquote></div><br>Yeah, the paper showed that a SMART failure indicates that your drive is highly likely to fail, so it is useful in that case. But not having any SMART failures doesn't mean that your drive is fine either. Basically there are lots of ways the drive can fail and many of them are not detected by the SMART tests.<br>