Re: [Freeipmi-devel] Re: ipmi sensor thresholds

[Albert Chu]

> So it seems that since the fan speed sensor is the only one with an 
> undefined upper critical threshold, you have to treat this sensor as a 
> special case.  

I don't think this needs to be handled by a special case.  The threshold
checking algorithm should be able to work around those values that
aren't defined.  My pseudo-code thinking is along the lines of:

if (reading > nominal)
{
   if (non-critical is defined && reading < non-critical)
       ouptut warning and return
   if (non-critical and critical is defined 
       && reading > critical
       && reading < non-critical)
       output critical and return
    else if (non-critical undefined
             && critical defined
             && reading < critical)
       output warning or alert or whatever;
   /* similar for non-recoverable and sensor min/max, but probably
      deeper and more complicated */
   
   if (nothing is defined except for nominal min/max)
      output unspecified fault
}

The above is just to illustrate my thinking on this, hopefully it makes
sense.  Ab, Ian, is this logically how you're looking at things?  

Al

--
Albert Chu
address@hidden
Lawrence Livermore National Laboratory

----- Original Message -----
From: Robin Goldstone <address@hidden>
Date: Thursday, August 12, 2004 5:36 pm
Subject: Re: [Freeipmi-devel] Re: ipmi sensor thresholds

> On Thursday 12 August 2004 05:14 pm, Anand Babu wrote:
> > So the fields we need to look at are
> >    Lower non-recoverable threshold: 0.00 RPM
> >    Upper non-recoverable threshold: 0.00 RPM
> >    Lower Critical threshold: 2100.00 RPM
> >    Upper Critical threshold: 0.00 RPM
> >    Lower non-critical threshold: 0.00 RPM
> >    Upper non-critical threshold: 0.00 RPM
> >    Sensor reading: 2680.00 RPM
> >
> > FAN sensor has no [Non-critical] [Non-recoverable] states.
> > Possible states are [OK] [Critical] [Unspecified fault].
> >
> > ONLY POSSIBLE RULE IS IN THE ABOVE CASE IS:
> >  if (Sensor reading <= Lower Critical threshold "2100.00 RPM")
> >     [Critical]
> >  else
> >     [OK]
> >
> > CALCULATION FOR THE ABOVE CASE:
> > Because "Sensor reading" (2680.00 RPM) is above "Lower Critical
> > threshold" (2100.00 RPM), State is [OK].
> 
> So it seems that since the fan speed sensor is the only one with an 
> undefined 
> upper critical threshold, you have to treat this sensor as a 
> special case.  
> For all other sensors:
> 
> if (Sensor reading <= Lower critical threshold) || (Sensor reading 
> => Upper 
> critical threshold)
>   [Critical]
> else
>   [OK]
> 
> It seems silly to me for you to have to treat this one sensor as a 
> special 
> case in your code.  It would be better if we could just get Intel 
> to put a 
> value for the upper critical threshold.  In one sense, I understand 
> their 
> point that there is really no concern about a fan running too fast. 
> But in 
> reality, there has got to be some maximum physical RPM limit 
> associated with 
> the fan.  If they would just use that as the upper critical 
> threshold, then 
> you would not need a special case for sensor #33.
> 
> -Robin
> 
> 
> 
> 
> _______________________________________________
> Freeipmi-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/freeipmi-devel
>