On Tuesday 07 August 2001 01:01 pm, Michael Schwendt wrote:
> Does anyone know how to reach the project maintainers?
>
> Message from yahoo.com.
> Unable to deliver message to the following address(es).
> <[EMAIL PROTECTED]>:
> Sorry, I wasn't able to establish an SMTP connection. (#4.4.1)
Hrmm, that address was valid not too long ago, and is still listed on
their web site. I used it perhaps a month ago.
> Question 2:
>
> # rpm -q lm_sensors
> lm_sensors-2.5.5-3
>
> I think the naming of "*_over" and "*_hyst" is swapped:
>
> set temp1_hyst 50
> set temp1_over 52
> set temp2_hyst 55
> set temp2_over 52
> set temp3_hyst 60
> set temp3_over 65
These are confusing, but correct, I believe. I decided this after
having read the notes in the config file 30 or 40 times. :) An argument
could be made either way, when you take the sensors output below into
consideration. In this usage, for temp1, the overtemp alarm will turn
on at 52, and off at 50.
> SYS Temp: +48.3°C (limit = +50°C, hysteresis = +52°C)
I took this to mean that while 52 is the limit, to prevent oscillation
between alarm on and off, we alarm at 52, and remain in the alarm state
until the temp cools to 50.
If it were handled the opposite way (limit 52, hysteresis 50), the
alarm would remain on, after the temp had dropped below the the high
limit. You'd see (after the alarm had triggered at 52, and we cooled a
bit):
SYS Temp +51C (hysteresis= +50°C, limit = +52°C) ALARM
I'm sure there are those who would argue that we shouldn't be in an
alarm state if the temp is below the limit.
For those who haven't fooled around with lm_sensors, the config file
sayeth:
# Set your temp limits here. Remember, 'tempX_over' is the temp at
# which an alarm is triggered, and 'tempX_hyst' is the temp at which a
# alarm turns off.
# Setting tempX_hyst to a few degrees below the corresponding tempX_over
# prevents an oscillation between alarm on and off states. This kind of
# oscillation is known as hyteresis, thus the name. (You typically get
# the most serious and troublesome hysteresis when a sensor triggers
# something to reduce the temp, thus creating a negative feedback loop.
# Even without that, we would still get some oscillation when the temp
# hovers around the limit due to noise.)
-D (Devil's Advocate?) ;)
_______________________________________________
Seawolf-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/seawolf-list
