[mrtg] Looking for some advice on the gaps we get in MRTG graph at peak traffic

[McDonald, Dan]

From: Greg Cory <gscjunk at gcmuni.net<mailto:gscjunk at gcmuni.net>>
Date: Monday, May 19, 2014 at 11:41 AM
To: "mrtg at lists.oetiker.ch<mailto:mrtg at lists.oetiker.ch>" <mrtg at lists.oetiker.ch<mailto:mrtg at lists.oetiker.ch>>
Subject: [mrtg] Looking for some advice on the gaps we get in MRTG graph at peak traffic

We have mrtg set up and running, it’s been working good till a few months ago as our traffic increased.

Now during the peak of our traffic the mrtg graphs get some gaps in them.  Usually ever night around 19:00 to 22:00.

Probably when backups run….

I would guess that you are using snmp v1 instead of snmpv2 or v3.  Thus, you have 32-bit counters instead of 64bit.  In a 5 minute period they roll over twice when you exceed about 130mb/s.  The solutions are to switch to 64-bit counters (by changing the snmp version to 2 or setting up the whole snmp v3 thing). Or change the polling interval from 5 minutes to 1 minute , which would let you poll a circuit with about 620mb/s before it would roll over twice between 1-minute polls.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/mrtg/attachments/20140519/9a99852a/attachment.htm