LMS 2.6 - DFM Alerting false positives for modules

tadams7541 · ‎07-30-2010

Hi,

I'm having an interesting problem that I thought you guys could help me with. I'm running LMS 2.6 with DFM 2.0.13. About once or twice every couple of weeks, I will see an alarms hit DFM saying that a bunch of cards are missing from switches. It will basically tell me that the cards no longer exist in the switch, but when I log into the switches to check on them, the cards are all status "OK." If I do a rediscovery on the switches, the alarms will eventually disappear. Initially I thought that this was due to an ifindex change in the SNMP information, but we have "ifindex persist" set on all of the switches, so the ifindex numbers shouldn't change.

Has anybody else seen anything like this before, or know of a way I can either fix or workaround it? It's getting frustrating because I'm constantly having to let people know that these are false alarms and not to worry. It's creating a "boy who cried wolf" situation, as people may not trust DFM when they see that there is an alarm and then be slower to act on real problems.

Thanks for any help you can provide. These alarms are making me nuts!

Joe Clarke · ‎07-30-2010

The only issue of which I am aware which could account for a problem like this is related to VSS management ( CSCta08882). However, given you're on DFM 2.0, I don't think that is the case.

This could be difficult to troubleshoot. The problem may be on the device side. That is, the device may be returning the wrong status for m

odules from time to time (or not returning certain objects). To determine this, a sniffer trace of SNMP traffic would be required at the time DFM detects this false NOTPRESENT situation.

If the problem is on the DFM side (meaning the sniffer trace shows that the switches are returning correct data), then debugging needs to be applied to the DFM server. This is an involved process, so I would recommend you open a TAC service request.