# Monitor external programs via SNMP (v. 1-3)
# (based on netsnmp-freespace.monitor)
#
# Usage: 
#       [-h]                # Usage
#       [-t Timeout]        # Timeout in ms (default: 1000000)
#       [-r Retries]        # Retries before failure (default: 5)
#       [-v SNMPversion]    # 1,2 or 3 (default: 1)
#       [-c Community]      # For SNMP v.2 (default: public)
#       [-u snmpuser]       # For SNMP v.3 (default: initial)
#       [-l seclevel]       # For SNMP v.3 (default: noAuthNoPriv)
#       [-A authpassphrase] # For SNMP v.3 
#       [-n num[,num...]]   # Zero-indexed external program number(s); programs
#                           # are numbered by the order they appear in the
#	                    # remote snmpd.conf. If monitoring specific programs
#	                    # on multiple hosts, they must be consistent!
#	                    # Default is to monitor all.
#       host [host ...]
#
# This script monitors one or more external programs run by the UCD
# SNMP agent. Specific programs to monitor can be specified with the
# "-n" option; these are zero-indexed in the order they appear in the
# monitored host's snmpd.conf file. Default is to monitor all.
#
# The summary output line will be of the form "host:name[,host:name]"
# where "name" is the name of the failing program (the "extNames"
# field as defined in snmpd.conf; not the path to the program). The
# detail lines will contain full error text from the failing program
# and the error value it returned.
#
# The script will exit with 0 value 1 for an extNames program failure
# and 2 for an SNMP error.
#
# BUGS AND LIMITATIONS: This is designed to handle programs that only
# return one line of output via snmpd; that is, with simple programs
# run via the "sh" or "exec" directives in the snmpd.conf file and NOT
# with programs run by "exec" and returning data in their own MIB
# tables. Actually, I've only gotten the "sh" directive to work with
# ucd-snmp-4.2.1 under Solaris. Also note that when given an external
# program number that doesn't exist on the monitored host, the script
# will return the output for program number 0 and will not report an
# error. In some situations (e.g. sending v. 1 request to a host
# configured only to respond to v. 3) the script will fail silently,
# because the SNMP module doesn't report an error.
#