Inaccurate Interface Polling statistics in Cisco 6509

Question

Hi all,

i need some help troubleshooting why NPM is reporting inacurate interface utilisation on a switchport on our Cisco 6509 switch.

I have been monitoring the switch interface stats through the console of the switch and the utilisation has never dropped below 110000000 bits/sec however the NPM graph and statistics are showing the utilisation fluctuate between 120mbps & 33mbps...

I have recently moved polling responsibility to an additional polling engine. Do i need to do some tuning or something? I have never touched the global polling settings so all the timers etc are the same.

Can anyone help?

Let me know if you want me to provide any more information.

thanks

Mario De Rosa

FormerMember · Answer

Check to see if SNMP is running up the CPU on your SUP.

If so, and if you have a very large ARP table on your 6506, you will need to drop L3 topology polling or setup a view in your SNMP config on that switch to filter out ARP table scans via SNMP.

Check this KB From Cisco: Catalyst 6500/6000 Switch High CPU Utilization - Cisco Systems

And this previous Thwack thread: