The explanation i've heard (and I think it does make a little bit
of sense but is mostly a crock, if you have it, put the right
value in it) is that if one has an interface that isn't fast enough
to roll in a reasonable amount of time (t1, ds0, bri, etc..) the
64-bit counters are not used and one should use the 32 bit counters.
i've found this to not be an absolute with cisco (not too shocking)
if you have a 3600 with a FE interface that can roll that fast.
Cisco needs to organize their snmp code such that this
will no longer be a problem. It's tiresome to collect stats
which don't mean anything, lose a few days of data while
tracking down that router X doesn't return usable values in
its 64-bit counters, while router Y does.
- Jared
On Tue, Apr 16, 2002 at 05:17:38PM +0200, Gert Doering wrote:
> hi,
>
> On Mon, Apr 15, 2002 at 05:12:27PM -0700, Zach Wilkinson wrote:
> > As you can see, I now get entries for the ATM interfaces but the counters
> > remain at zero. Is there any additional configuration to enable the
> > counters?
>
> No, this is just IOS as we all know and love it.
>
> My application that uses Counter64 falls automatically back to Counter32
> if ( both in/out Counter64 == 0 and Counter32 != 0 )...
>
> <rant>
> Cisco and interface counters is a major PITA. No matter which counters,
> SNMP or "show int", 5 minute average or octet counters, all of them are
> broken in funny and interesting ways.
> </rant>
>
> gert
>
> --
> USENET is *not* the non-clickable part of WWW!
> //www.muc.de/~gert/
> Gert Doering - Munich, Germany gert@greenie.muc.de
> fax: +49-89-35655025 gert.doering@physik.tu-muenchen.de
-- Jared Mauch | pgp key available via finger from jared@puck.nether.net clue++; | http://puck.nether.net/~jared/ My statements are only mine.
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:41 EDT