Question about smartd temp. readings.

[Björn Persson]

Erik P. Olsen wrote:
> I've noticed in logwatch the following smartd readings of hard disk
> temperature:
>
>  /dev/sda :
>     Usage: Temperature_Celsius (194) changed to
>       171, 161, 157, 161, 157, 161, 157, 152, 157, 152, 157, 161,
>       157, 152, 157, 152, 157, 152, 157,
>
>  /dev/sdb :
>     Usage: Temperature_Celsius (194) changed to
>       157, 148, 144, 141, 144, 141, 144, 141,
>
> Is this really Celcius degrees? To me it looks more like Fahrenheit. If it
> indeed is Celcius it won't take long before the disks melt down :-(

No, that's not degrees Celsius. SMART attributes use normalized values. The 
closer to zero an attribute gets, the more likely it is that the disk will 
fail soon. Common practice seems to be to use 100 for "working normally" and 
200 for "has never happened". For the temperature I'd assume that 100 means 
normal working temperature and a value greater than that is cool.

Run "smartctl -A /dev/sda" and look in the column RAW_VALUE. That should be 
the temperature in degrees Celsius. Otherwise the attribute shouldn't be 
called "Temperature_Celsius".

In my opinion it's rather silly to alert you every time the temperature 
fluctuates a little. I've put the line "/dev/sda -d sat -H -m root" in my 
smartd.conf. That way it will send me email if the disk is about to fail and 
not bug me otherwise. (Replace "sat" with your disk type.)

Björn Persson
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
URL: <http://listman.redhat.com/archives/fedora-list/attachments/20081117/bc6b256f/attachment-0001.sig>