![](https://secure.gravatar.com/avatar/a4042920f4bf89a219241c65ae64c5d8.jpg?s=120&d=mm&r=g)
I had the same issue with the 2960S switches, polling the cisco-cef and fdb-table modules would extend our polling time out to exorbitant amounts of time for those devices. Disabling those two modules, my polling time is now around 7-8 seconds for each device.
-----Original Message----- From: observium [mailto:observium-bounces@observium.org] On Behalf Of Richard Kuhns Sent: Wednesday, January 28, 2015 2:52 PM To: observium@observium.org Subject: [Observium] An interesting issue with Cisco 2960Cs...
Hi all,
This is just on the order of a 'watch out' notice.
We're currently on 0.15.1.6206, but had the same issue with earlier versions and with the CE.
We have 5 Cisco 2960Cs that we monitor, and all were taking horrible amounts of time to poll. Always at least 200 seconds, and often over 300.
After some experiment, we discovered that they apparently don't care to talk about CEF (Cisco Express Forwarding). If Observium asks them about it, the CPU usage climbs above 60% and stays there for quite a while.
If I go to the device --> edit --> Modules and disable cicso-cef, everything is fine. The CPU drops under 5%, and the poll time is 5 seconds or less.
I'm pretty sure we're not losing any useful info, since I don't believe the 2960Cs support CEF in the first place.
As I said, this is just a heads up.
Observium is a great tool.
Thanks!