Nearly for 6 days i 'm trying to get over this problem .
I have 2 machines . One of them is manager 10.201 machine, and the other is agent (10.226)
i've configured manager's snmptrap.conf file like this .
authCommunity execute public default .1
traphandle default /usr/bin/traptoemail -s localhost -f snmp@localhost root@localhost
and then i've configured agent's snmpd.conf file like this.
###########################################################################
# SECTION: Access Control Setup
rocommunity public
rwcommunity private
###########################################################################
# SECTION: System Information Setup
syslocation "Izmir, 226. machine "
syscontact [email protected]
###########################################################################
# SECTION: Trap Destinations
trap2sink 10.0.0.201 162
authuser read -s v2c guest_user noauth .1
authuser read -s usm guest_user noauth .1
authcommunity read public default .1
iquerySecName guest_user
agentSecName guest_user
monitor -u guest_user -r 1 "interface down" -o ifDescr ifOperStatus != 1
But I want to trap for agent machine's CPUIdle rate. For example I can get CpuIdle like this way by snmpwalk command.
# snmpwalk -v 2c -c public 10.0.0.226 .iso.org.dod.internet.private.enterprises.ucdavis.systemStats.**ssCpuIdle**.0
* I want to trap to manager host(10.0.0.201) when agent's (10.0.0.226) *ssCPuIdle rate is over 98.*
Is there anyone can help me ? Where I ve done mistakes,conf files or anything else ?
One thing to note is that the ssCpuIdle variable is actually going away. It was measured over a fixed period of time, which is a problem as it's inaccurate.
If you still want to use it on an old version of snmpd, then the following monitor line will probably do what you want:
monitor -u guest_user -r 1 "idle too high" ssCpuIdle > 97
If you use the newer ssCpuRawIdle object, then you'll need to monitor it's delta instead:
monitor -u guest_user -r 1 "idle too high" -D ssCpuRawIdle > NUMBER
But the size of the NUMBER will likely depend on the system in question (specifically, how many cores). You can use the description of the object to come up with an initial calculated guess and go from there:
DESCRIPTION "The number of 'ticks' (typically 1/100s) spent idle.
On a multi-processor system, the 'ssCpuRaw*'
counters are cumulative over all CPUs, so their
sum will typically be N*100 (for N processors)."