I have seen similar results where nothing appears to be happening with traps, any fault conditions are detected via the periodic polling mechanism, not dynamically via trap handling.
Cisco have suggested altering the log window update frequency; Extract from message train of 'Ask the Expert Forum' follows :
Calvin,
Thanks for the reply.
Yes it's a good point about the message console display frequency as DFM may be detecting events but not displaying them, although I would expect an event to pop up as soon as it is processed.
Thanks for the URL, I had already seen the Apppendix in the user guide with the defined traps.
I will deal with any similar issues in the future on a case by case with the TAC.
Ian.
>>> Calvin Chai <cwc@cisco.com> 20/02/02 17:28:46 >>>
Hi Ian,
You are correct - if you pulled out a network connection, DFM should
generate a notification for that since it acts upon the standard link
up/down traps. One possibility is that the refresh frequency of your
monitoring console is set to update at a lower frequency. In any case,
provided you have a CCO login, you can access the DFM user guide which
points you to which traps that the application listens for:
http://www.cisco.com/univercd/cc/td/doc/product/rtrmgmt/cw2000/dfm/dfm1_0/user/appb.htm
Hope this helps.
- Calvin
At 04:47 PM 2/19/2002 -0800, David Stiff wrote:
>Calvin, Can you provide an explanation to this customer about his DFM
>questions and possibly a like to the DFM supported traps and any sort of
>white paper that might help him.
>
>Thanks - David
>
>At 04:42 PM 2/19/2002 +0000, Ian Swift wrote:
>>David,
>>
>>Unfortuntely I missed your forum at CCO's Net Pro centre.
>>
>>Do you have time to expand on one of your points about DFM, or pass this
>>on to a colleague who does.
>>
>>I don't want to solve a specific problem but understand what DFM 'should'
>>be doing.
>>
>>Your statement was :
>>¯----------------------------------
>> dstiff - CISCO SYSTEMS
>>
>>Feb 7, 2002, 10:53am Pacific (8.1.1.1.1.1.1)
>>I resent the lab docs to you. It turns out the attatchment was too large
>>so I zipped it this time.
>>
>>Answer) DFM is not a generic SNMP trap reciever. It only acts on a
>>limited set of traps for events it is monitoring. It is mainly a MIB
>>poller. To deal with the SNMP traps from IPM you need to use a separate
>>tool to receive and act upon them. There are freeware tools for this as
>>well as commercial tools such as HP OpenView. You can point your traps to
>>DFM and have it forward them to another system for processing. Then you
>>can act upon the IPM traps as well as others that DFM doesn't worry about.
>>
>>- David
>>¯--------------------------------------------------
>>
>>I am not concerned with the IPM forwarding aspect of your questioner, but
>>'vanilla' traps coming from network devices.
>>
>>I have implemented several systems with LMS and notice the same behaviour
>>with all of them, namely DFM doesn't do anything with traps, it seems to
>>be MIB polling only.
>>
>>The default device polling time is 4 mins. and I have seen fault
>>conditions detected within a 4 minute window. (presumably by polling the
>>device)
>>If however you have a Cisco device set with a trap receiver address of
>>the LMS platform, nothing happens say, if you pull a Gig link connecting
>>a neighbouring switch. At the very least I would expect 'Link Down' to
>>trigger an alarm, let alone Spanning Tree alarms, before the 4 minute
>>window is up.
>>Similarly for other created fault conditions.
>>
>>Is my understanding correct, can you provide any hints or links to assist
>>here ?
>>
>>Many thanks,
>>
>>Ian
