The event rule here from HaydnH is awesome and works a treat. However, I think an easier way to create and process events based on net-snmp monitoring processes (or logfiles) is to use the -e EVENT option in the monitor statement in your snmpd.conf file. That way, you can control the exact OID of the event that is sent, rather than relying on a standard event from the DisMan MIB. So, my monitor command would be:
> monitor -u userX -r 1 -D -S -e myEvent -o prErrorFlag -o prNames -o prMin -o > prMax -o prCount -o prErrMessage "procWatch" prErrorFlag != 0 > notificationEvent myEvent .1.3.6.1.4.1.1234.123 > > The notificationEvent statement defines the OID of the event you want to send - the varbinds from the monitor statement are appended. Thus, you only need to check the varbind for prErrorFlag to see if it is a "good news" event or a "bad news" event. You can create discrete events, at source, for each failed process, (with a different event OID) which gets rid of all the fancy processing to determine the instance of the process table that has caused problems. Cheers, Jane -------------------- m2f -------------------- Read this topic online here: http://forums.zenoss.com/viewtopic.php?p=33000#33000 -------------------- m2f -------------------- _______________________________________________ zenoss-users mailing list [email protected] http://lists.zenoss.org/mailman/listinfo/zenoss-users
