Actually, snmpCollect knows, and writes it down in the snmpCol.trace file
and I have always wished it would report it more loudly. Check the trace
file and you will see a report that the node is not responding to snmp, and
collection on that node is deferred, for an hour by default (I usually knock
that down to 15 minutes).  You might try monitoring that file for the words
in the message, maybe with a Tivoli logfile adapter, if you are a full
Tivoli shop, or with the SIA and MLM/APM. The SIA has the ability to
report on the presence or absence of strings in a file. Then you would
receive a trap when snmpCollect noticed the agent was down.

Cordially,

Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking



VInce;

I know of no way to get NetView to generate a trap when a device looses its
ability to talk SNMP.  I suppose you can write a script to periodically try
an snmpget on a device and then manually send a trap to NetView when a
failure occurs.

Art DeBuigny
debuigny AT dallas DOT net
art.debuigny AT bankofamerica DOT com
Bank of America Network Operations
-----Original Message-----
From: Vince Rosso <Vince_Rosso AT TRANSALTA DOT COM>
To: NV-L AT UCSBVM.UCSB DOT EDU <NV-L AT UCSBVM.UCSB DOT EDU>
Date: Wednesday, July 21, 1999 12:01 PM
Subject: SNMP Monitoring


>Vince Rosso@TRANSALTA
>21/07/99 10:59 AM
>
>We are running  v5.1.1 Netview on Digital UNIX v4.0d.  We are trying to
>find a way to check if an SNMP agent is running on a device.   Often times
>the device is pingable but the SNMP agent has failed which doesn't allow
>for data collection or trap generation from the device.   Is there a way
>for Netview to check for SNMP running on  nodes and if it isn't then
>generate a trap so that it can be dealt with.  Thanks!