[mrtg] Measuring (Wan-Out) - (Lan-In) is not working

Feijo, Fernando Fernando.Feijo at aaronrents.com
Mon Oct 20 06:05:15 CEST 2008


I was wondering if someone could figure out what is going wrong on this
one.

 

We have a VPN concentrator anchoring about 1,600 sites. MRTG measures
successfully both WAN and LAN interfaces.  A lot of the sites in the
concentrator mesh to other sites also anchored in the box. 

 

We wanted to measure the "bounce" traffic that never enters the LAN,
just bounces off the concentrator in between VPN sites. The solution we
came up to was to subtract from the WAN-out (SNMP 2) the LAN-in (SNMP
4).  That would define the bounce: traffic leaving the concentrator
towards the cloud that did not come from the LAN.

 

The target line attempted was short:

 

Target [sonicw-3rd-bounce]:     (2:string at 172.16.9.226::10:10) -
(-4:string at 172.16.9.226::10:10)

 

 

This delivered garbage numbers unrelated to actual traffic. We thought
that maybe the (IN/OUT swap on using the minus sign in the LAN (4)
interface was behind the problem, and tried directly with OIDs:

 

Target [sonicw-3rd-bounce]: (
(.1.3.6.1.2.1.2.2.1.16.2&.1.3.6.1.2.1.2.2.1.16.2:string at 172.16.9.226::10
:10) -
(.1.3.6.1.2.1.2.2.1.10.4&.1.3.6.1.2.1.2.2.1.10.4:string at 172.16.9.226::10
:10) ) 

 

Still dreck results. Attempts to play with the parenthesis, adding more,
did not help. 

 

One strange thing is that the garbage results are unrelated to the
arithmetic possible with the four combinations of WAN/LAN and In/Out.
For instance, if traffic is hovering around 4 megs in the LAN and 6 in
the WAN, we could get 26 megs on the target above.

 

Thanks if someone can point in the right direction.

 

Fernando

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/mrtg/attachments/20081020/6868c6a5/attachment.html 


More information about the mrtg mailing list