Franc Zabkar wrote:
> In the following example ...
> ... which appears on the Help page ...
This is a rather old example, I think from (perhaps) six years ago.
Since that time I think Maxtor has fixed some firmware bugs in their
lifetime counters, and smartmontools has gotten somewhat better at
interpreting the attribute values.
See the FAQ page http://smartmontools.sourceforge.net/faq.html and
search for 'Maxtor'.
> ... I notice that the Power_On_Hours count is 32671.
> The cooked value has fallen from an initial value of 253 to a current
> value of 236, suggesting that the lifetime of the drive is ...
> 253 / (253 - 236) * 32671 = 486221 hours
> This equates to 55.5 years.
> Surely this does not make any sense.
You are assuming (incorrectly) that the normalized Attribute values are
linear functions of the raw Attribute values. This is often not correct.
> Furthermore, the Start/Stop count is 62 which would suggest that the
> drive runs for an average of 7842 hours between power cycles. That's
> about 327 days.
> If OTOH the attribute were measured in minutes, then the lifetime
> would be about 1 year and the drive would be power cycled every 5
> days. Does this make sense, ie 9-5, 24/7, off on weekends?
Interpreting the raw attribute value as minutes does not make sense,
since the self test log shows that the device is more than 5000 hours
old. Hence the power on time would have to exceed 300,000 minutes.
If you want to know how old a disk is, a fairly reliable way is to run a
short self-test, then look in the self-test log for the age. However
some disks and vendors have bugs in this part of the SMART firmware (see
FAQ above) so this is not 100% reliable. But better than 99%!