We're running ntop 2.2 and are trying to switch from using a monitor
port to having our routers export NetFlow statistics. I'm running into a
number of problems. As near as I can tell, ntop locks up after some
period of time in two cases: 1) If we're doing running with both NetFlow
and the high-traffic mirror port; session track is disabled. 2) If we
run with NetFlow only and enable session tracking.

I *think* things are okay with NetFlow only if session tracking is
disabled.

I tried upgrading to ntop 3.3 and using just NetFlow (with session
tracking disabled). It seems to separate the interfaces coming in from
NetFlow (on the hosts page). With ntop 3.2, the CPU load was less than
20% at all times. With ntop 3.3, the CPU is maxed out immediately and
stays that way. Could this be due to the change in interface handling?
We have a LOT (a couple thousand) sub-interfaces on one router.

Is there some way (either on the Cisco routers or in ntop) that we can
get the old behavior back to see if that's the cause of the CPU
problems?

Thanks,
Richard

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to