That's odd -- what settings are you using in MRTG?  Are you using native mode 
or RRD mode?  Is the reset based on time (never head of this before, but try 
the solution below), counter size (if it is wrapping, use 64bit counters and 
SNMPv2), or polling (use 'absolute' datatype if it resets on read)?

I would have expected MRTG/RRD to simply log an UNKN immediately after the 
reset as it would appear to be like a counter wrap but the resulting rate would 
be > MaxBytes and so discarded.  It would still mean you lose 1 in 3 samples, 
if you're polling every 5 min.

Maybe poll every 1 min, using a COUNTER datatype, and set MaxBytes 
appropriately so the rest is detected as a counter wrap and discarded?  Also, 
if you set it to use the previous value on unknowns (default behaviour for 
native mode or 14all, extended option unknasprev on MRTG/RRD/Routers2) you'll 
hide the unknowns.

Steve


Steve Shipway
University of Auckland ITS
UNIX Systems Design Lead
[email protected]<mailto:[email protected]>
Ph: +64 9 373 7599 ext 86487

________________________________
From: [email protected] 
[[email protected]] on behalf of pradip 
Bhojania [[email protected]]
Sent: Wednesday, 28 December 2011 7:48 p.m.
To: [email protected]
Subject: [mrtg] Plotting counters which reset after every 15mins.

The ethernet interfaces which i am monitoring resets the counters every 15min. 
MRTG in this case plots a graph which takes a dip after every 15mins. How do i 
configure mrtg to add these counters(after every 15min) to the previously read 
counters and plot the same. Any help is really appreciable.

--
Regards,
-Pradip Bhojania
_______________________________________________
mrtg mailing list
[email protected]
https://lists.oetiker.ch/cgi-bin/listinfo/mrtg

Reply via email to