is my system drive getting too hot?

[Tom Metro]

Stephen Adler wrote:
> ...the temperatures over 100. Is this something I need to worry about?
>   /dev/sda :
>      Usage: Temperature_Celsius (194) changed to
>        114, 113, 114,

In my experience, SMART is great for reporting the pass/fail conditions, 
but near useless for the quantitative attributes, like temperature, 
because it seems there is no standard for how to report those numbers. 
(Also watch out for attributes that go down instead of up to indicate a 
worsening situation.) I've seen some recommendations to use the "raw" 
values returned by those attributes (smartd has options for that). I 
guess this is because smartd is applying a default conversion algorithm 
that is incorrect for the particular drive.

In theory this is something that should be addressed by smartd's drive 
database, but it doesn't seem to work all that great in practice.

The "finger test" idea is probably a good recommendation. You can then 
baseline your drives and adjust the thresholds accordingly so you don't 
get emails when the drives are within normal range.

I'm currently using these temperature related options to smartd on a mix 
of WD and Seagate drives:

# -I 194 - ignore normalized Temperature
# -W 4,45,55 - track Temperature changes >= 4 Celsius,
#              report Temperatures >= 45 Celsius,
#              Send mail on SMART failures or when Temperature is >= 55 
Celsius.

and it seems to trigger failure notices only when there is a legit 
overheating problem.


> Also, I've had a lot of problems with WD drives in the past and 
> have sworn never to buy them...

If you blacklist a drive vendor every time you run into problems, you'll 
soon run out of vendors. :-)

  -Tom

-- 
Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile: http://tmetro.venturelogic.com/