lm_sensors output

Curt Moore jcmoore at atipa.com
Thu Oct 2 13:56:04 EDT 2003


This is really the bad thing about lm_sensors which some have touched on
previously; too much guesswork.  Many times even if the drivers are
present and up to date for your specific hardware, the values may be
meaningless as different board manufacturers may choose to physically
connect the monitoring chip(s) to different onboard devices, such as in
the case with fans.  You have to have a knowledge of which onboard piece
of hardware is connected to which input of the monitoring chip in order
to make sense of the sensors output.  Don't get me wrong, when
lm_sensors works, it works great but sometimes it takes a little work to
get to that point.

Even if the values are sane for your hardware, you still have to go into
the sensors.conf and set max, min, and hysteresis values, if you so
choose, in order to have this information make sense for your specific
hardware.

In recent months, vendors such as Tyan have begun to distribute
customized sensors.conf files for their boards which take into account
the differences between boards and how sensor chips are connected to the
onboard devices for each of their boards.

As Don mentioned earlier, IPMI is more generalized and is much easier to
ask for "CPU 1 Temperature" and actually get "CPU 1 Temperature" instead
of data from some other onboard thermistor.  A mistake in this area
could end up costing time and money if something overheats and it's not
detected because of polling the wrong data.

>From my experience, it would be very difficult to come up with a
generalized set of sensors values to work across differing motherboard
types.  A "standard" such as IPMI makes things much easier to accurately
collect and act upon as all of the "hard" work has already been done by
those implementing IPMI on the hardware.  One would hope that these
individuals would have the in-depth knowledge of exactly which values to
map to which sensor inputs and any computations needed for these values
so that clean and accurate values are returned when the hardware is
polled.

-Curt

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Curt Moore
Systems Integration Engineer
Atípa Technologies
jcmoore at atipa.com (O) 785-813-0312 (Fax) 785-841-1809
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list