> From owner-cisco-nsp@nic.iagnet.net Fri Jan 16 08:07:26 1998
> From: "Y.L. Yeng" <ylyeng@cuhk.edu.hk>
Sorry, this is an old thread. I have noticed something weird, though.
> At 11:26 PM 1/15/98 -0500, you wrote:
>
> >
> >Currently developing a system to track usage stats on cisco interfaces in
> >order to bill based on usage (burstable service).
> >
> >Pulling the IfOctets of total bytes, the stats seems to mysteriously reset
> >to 0 on occasion. No reboots, etc.. throws the count off and makes one
> >wonder how reliable the 'total bytes' stats are.
> >
> >What is the best method of getting reliable cisco usage stats, on a 5
> >minute basis, over a long period. Prefer SNMP because of the low resource
> >overhead vs. an expect script. Unable to find a SNMP variable that shows
> >the 5-minute average as seen in 'sh int'.
> >
> > Stb
> >
> >
>
> MIB object ID .1.3.6.1.4.1.9.2.2.1.1.6 and .1.3.6.1.4.1.9.2.2.1.1.8 should
> give the 5-minute average of input rate and output rate in bits per second.
If I query the router too often (e.g. every minute), I get some quite strange
values, like the 5 minutes average speed is _higher_ than the speed of the
interface (although the real throughput, measured with If{In|Out}Octets is
much less: let's say one fourth of what I get with locIf{In|Out}BitsSec). :(
Is this due to the fact that I query too often? What makes me even more
confused is that a "sh int" gives 'acceptable' values, even while I query
the router for locIf{In|Out}BitsSec every minute...
Did I make some mistake? If not, has anyone seen this before? Is this a known
problem? I apologize if it is either known or I made an error while
investigating.
Regards,
Janos Zsako
PS. I have tested this on a 7010 with 1 FSIP (8 serial) running IOS 10.3(7).
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:15 EDT