[mrtg] Re: MRTG General
jeffl at comix.santa-cruz.ca.us
Thu Oct 25 01:38:19 MEST 2001
On Wed, Oct 24, 2001 at 04:15:59PM -0500, Tony Andrews wrote:
> In this case, I have an Extreme Networks Summit48 and I want to watch
> traffic on two ports that are connected to DSL connections. Through the
> web interface on the switch I can see what the maximum/minimum input and
> output rates are, but I would like to see some traffic patterns over
> time... With a very basic MRTG setup, I am getting readings that are well
> below what the switch appears to be handling in real-time... for example
> one graph shows a maximum of 64K input, but when I look at the statistics
> on the switches web interface I see that 200 k is probably a more likely
Hint: k=kbits/sec, K=KBytes/sec. At least be consistant.
> Any suggestions as to how I could collect more accurate information for
> these types of settings?
You probably don't have an "accuracy" problem.
MRTG drops any values that are over the:
value. If SNMP returns a value over this limit, it enters
a zero in the corresponding log file.
Most of the fast switches can belch numbers very quickly. They
will "wrap" a 32 bit counter in the 5 minute polling interval.
Use a MIB browser to see if the counters are wrapping. If
ifHC___ 64 bit counters are available, see the Configuration
Guide under "SNMPv2c" for the proper Target line configuration.
> I'm currently running "Mr. Daemon" to collect information... would polling
> more frequently provide closer data?
No. Getting wrong numbers more frequently will not enhance
accuracy. Use your MIB browser to check for what values are
actually returned and compare those with what ends up in
the MRTG logfiles. If it wraps or goes over MaxBytes, change
the Target line, MaxBytes line, or both.
# Jeff Liebermann Liebermann Design 150 Felker St #D Santa Cruz CA 95060
# 831.336.2558 voice 831.426.1240 fax 831.421.6491 digital_pager
# jeffl at comix.santa-cruz.ca.us http://www.cruzio.com/~jeffl
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
FAQ http://faq.mrtg.org Homepage http://www.mrtg.org
More information about the mrtg