[c-nsp] 5s or 10s periodic IOS tasks

Rubens Kuhl rubensk at gmail.com
Sat Jun 11 19:37:54 EDT 2011


On Sat, Jun 11, 2011 at 1:44 PM, Pete Lumbis <alumbis at gmail.com> wrote:
> Not sure on the timers but generally testing CEF/PXF switched issues
> (through the box) with process switched traffic (to the box) isn't
> always a valid check. I'd suggest pinging something directly connected
> to the 7300 so the traffic is CEF switched through the 7300 when
> testing.

The multicast traffic was going thru the box with similar loss patterns.

> If you are seeing packet loss to the box my guess would be interface
> issues or CPU issues. Do you see any drops on the interface or
> controllers?

After noticing a high amount of flushes, disabling spd (selective
packet discard) seemed to solve the packet drops, but a few more
business days will confirm if it was simply that. Frequency of Hello
OSPF packets (10s) was the master clock of the packet drops.

Rubens


More information about the cisco-nsp mailing list