[mrtg] Re: Inaccurate Peaks

John Haggerty haggerty at bnl.gov
Thu May 4 02:07:08 MEST 2006


This sort of thing happens to me now and then too, usually after a 
meltdown of some sort.  How can you get rid of the crazy peaks without 
blowing away all the log files?  (Just deleting numbers that look funny 
in the log files does not seem to accomplish this.)

Steve Shipway wrote:
>> On a under utilized STM-64 link, we observe many instances of 
>> inaccurate peaks in the graph. These peaks appear only on the 
>> IN-Traffic. When compared with the OUT traffic on the other 
>> end, except for the peaks the other points seem to match well.
>> We analysed the incoming SNMP packets using ethereal and did 
>> not see such large values. Has anyone observed such a 
>> behaviour before ? Also, are there any known issues like this 
>> that someone is aware of ? Thanks for the help.
> 
> Could be due to counter resets, and MRTG incorrectly assuming a counter
> wrap.  I'd suggest you use the AbsMax[] directive to force it to ignore
> values you know are way too high.  Also, check to see if the SNMP agent has
> been reset at corresponding times, or an interface reset, or something
> similar which could have caused the SNMP counter for that interface to reset
> to zero and confuse MRTG.
> 
> Steve
> 
> 
> --
> Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
> Archive     http://lists.ee.ethz.ch/mrtg
> FAQ         http://faq.mrtg.org    Homepage     http://www.mrtg.org
> WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi
> 

-- 
John Haggerty
email: haggerty at bnl.gov
voice/fax: 631 344 2286/4592
http://www.phenix.bnl.gov/~haggerty

--
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
Archive     http://lists.ee.ethz.ch/mrtg
FAQ         http://faq.mrtg.org    Homepage     http://www.mrtg.org
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi



More information about the mrtg mailing list