[Bigsister-general] Bug in CPU load test?

Hello all,

Digging in the mailing list´s  archives I have found that Matt Kettler reported (2005/02/22) that:

 “Using 0.99b2 I"m having trouble getting snmp based monitoring of a remote

 host"s CPU to work in a reasonable fashion.

 Let"s say I have a pair of example netware boxes, and localhost. netware1

 is a single CPU netware box, netware2 is a dual CPU netware box, and

 localhost is a single CPU linux box.

 So I create these DESCR statements:

 DESCR           features=local,unix,linux               localhost

 DESCR           features=remote,netware                netware1

 DESCR           features=remote,netware                 netware2

 Both netware boxes respond to snmpwalk for processor load:

 $snmpwalk netware1 public hrProcessorLoad

 host.hrDevice.hrProcessorTable.hrProcessorEntry.hrProcessorLoad.1 = 0

 $snmpwalk netware2 public hrProcessorLoad

 host.hrDevice.hrProcessorTable.hrProcessorEntry.hrProcessorLoad.1 = 30

 host.hrDevice.hrProcessorTable.hrProcessorEntry.hrProcessorLoad.2 = 3

 Now if I try to use "load" to monitor netware2 and netware1:

        netware1       load

        netware2       load

 I get the same load average for both boxes. Highly unlikely.

 Now if I try "cpuload":

        netware1       cpuload

        netware2       cpuload

 I get output for localhost"s "top" command for both. Including claiming

 they are running uxmon, and netware2 only shows 1 cpu. Clearly not right.”

I have exactly the same behavior using bigsister 0.99c3 and 1.02 : all my servers report the same load average, which is definitely wrong. Is this a problem of my bigsister configuration?, Does anybody knows how to solve it?

Thanks a lot.

Carlos