[mrtg] Re: question about SNMP on Server
Michael Votaw
mvotaw at telepath.com
Thu Apr 27 09:35:11 MEST 2000
There are a couple of things you should keep in mind. First of all, outside
of some unforeseen problem, MRTG in reporting exactly what it is getting
from your server. MRTG is probably looking in the MIBS and reading two of
them, ifInOctets and ifOutOctets. These are simply counters. For every
byte that goes in, ifInOctets is incremented. For every byte that goes out,
ifOutOctets in incremented. MRTG simply pulls this number once every five
minutes, finds the difference from last poll, and divides the total number
of octets sent and received in 5 minutes to give you the number of bytes per
second over a 5 minute period.
The representation of the "peaks" of utilization completely hinges on the
sampling rate.
You can even see this in MRTG right now by looking at the daily, weekly,
monthly, and yearly graphs to see the "peaks" of utilization dwindle into
lower averages. It is working exactly as it is supposed to. People have
often told me that they didn't want to miss peaks of utilization by not
polling frequently enough. If you were to take your sampling rate down to,
say, 1 millisecond, then anytime somebody sent one (1) packet of 1300 bytes
on an Ethernet segment, you'd see a 100% utilization for that time period.
Find out exactly how you are being billed. By total number of octets? By
"peaks"? Ask them if you can begin monitoring the server too for your own
records. You'll need the following info...
1)The IP address they are using to monitor the server
2)The community string they are using to get read access to MIBS
With these two items, you can begin monitoring the server for yourself. I
will tell you this, that MRTG is one of the more reliable polling utilities
I've seen. I have seen many others that were far more expensive, more
complex, and sometimes, didn't work as well. You are right to question the
data and ask to compare to something else. But from my experience, MRTG is
a simple tool with a small chance of reporting incorrectly(outside of system
failures).
> -----Original Message-----
> From: mrtg-bounce at list.ee.ethz.ch [mailto:mrtg-bounce at list.ee.ethz.ch]On
> Behalf Of Fragmeister
> Sent: Thursday, April 27, 2000 7:53 AM
> To: mrtg at list.ee.ethz.ch
> Subject: [mrtg] Re: question about SNMP on Server
>
>
> I am new to this and own a dedicated server which is hosted by a
> company who
> uses this meter and I have a few questions wondering if this meter is 100%
> accurate since I have been getting a very high bandwidth bill
> lately and can
> not see how my sites are doing all of this.. I looked over my stats and it
> does not match up to the amount of bandwidth my host is telling me and I
> want to know how good this program really is and if there is a chance it
> could be wrong?
>
> Has anyone ever got a false reading from it? Like anything on the market ,
> there is always that possibility of it not measuring properly and this may
> be that time, but only where I get stuck with the damn bill. I
> want to find
> out if there is something I can do on my end to watch this traffic and
> compare it to what my host is saying.. Can I install this meter without
> being at the routers? This way I can compare to what they show and see if
> they are bull crapping me..
>
> Anyway, your help is appreciated,
>
> Thanks!
>
> ___________________________________________
> John
>
> --
> Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
> Archive http://www.ee.ethz.ch/~slist/mrtg
> FAQ http://faq.mrtg.org Homepage http://www.mrtg.org
>
--
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
Archive http://www.ee.ethz.ch/~slist/mrtg
FAQ http://faq.mrtg.org Homepage http://www.mrtg.org
More information about the mrtg
mailing list