Re: Cache traffic - odd?

On Fri, 5 Sep 1997, Stephane Bortzmeyer wrote:

> We have exactly the same problem on <http://cache.cnrs.fr/MRTG/>. See the difference between the FDDI savings (8 % in bytes) and the Squid savings <http://cache.cnrs.fr/MRTG/stats/> (15 %).

What did you do to allow you to graph "Response time"? This is something
that could be very useful and I never thought to try to get MRTG to graph
it...

> >This has
> >to do with the way Ethernet handles packets (or so I'm told), and I don't
> >know of any way around it.
>
> I see no reason for that. Could you explain?

This is based on my observations of MRTG graphs on ethernet interfaces for
machines I administer. The numbers aren't correct. If you figure out the
*amount* of traffic sent or recieved (volume), it comes out right, but if
you comare it (the traffic on the ethernet interface) to the total traffic
on all of the routers, it isn't right. I trust the graphs for the router's
WAN interfaces. I also see the in and out being almost identical for the
Ethernet interfaces even though according to ifconfig it shouldn't be.
Someone else told me that it has to do with how Ethernet handles packets
on the MAC layer. If anyone has suggestions I would love to hear them. I
would like to have better graphs for my ethernet interfaces!

        -Bill
Received on Fri Sep 05 1997 - 11:55:22 MDT