NAME

nt-guide - The Windows NT Guide to MRTG 2.9.17


SYNOPSIS

Installing MRTG on an Windows NT or Windows 2000 box is not quite as click and point as some might want it to be. But then again it is not all that difficult if you follow the instructions below.


PREREQUISITS

To get MRTG to work on Windows NT you need the following:


INSTALLATION

I suggest you do the following from the machine that will be running MRTG, which, in this case is also a web server. All examples are for doing things to a LOCAL machine.

First
Unzip MRTG to C:\mrtg-2.9.17 on the WindowsNT machine of your choice.

Next
Install Perl on the same Windows NT machine. You might want to make sure that the Perl binary directory is listed in your system path.
 C:\Perl\bin;%SystemRoot%\system32;%SystemRoot%;...

You can manually check this by going to [Control Panel]->[System]->[Environment]

To see if everything is installed properly you can open a Command Shell and go into c:\mrtg-2.9.17\bin. Type:

 perl mrtg

This should give you a friendly error message complaining about the missing mrtg configuration file. Now, you have successfully installed mrtg and perl.


CONFIGURING MRTG

Now it is time to walk create a configuration for mrtg. But before we begin you need to know a few things, take an opportunity to gather the following information:

For the rest of this document we will be using device 10.10.10.1 ( a CISCO Catalyst 5000) with Community string public.  We are interested in monitoring traffic, and the CPU Load. Let's begin.

The first thing we do in setting up mrtg is by making a default config file. Get to a cmd prompt and change to the c:\mrtg-2.9.17\bin directory. Type the following command:

 perl cfgmaker public@10.10.10.1 --global "WorkDir: c:\www\mrtg" --output mrtg.cfg

This creates an initial MRTG config file for you. Note that in this file all interfaces of your router will be stored by number. Unfortunately these numbers are likely to change when ever you reconfigure your router. So in order to work around this you can get cfgmaker to produce a configuration which is based on Ip numbers, or even Interface Descriptions. Check the cfgmaker manpage

If you get an error message complaining about no such name or no response, your community name is probably wrong.

Now, lets take a look at the mrtg.cfg file that was created.

In Perl a # is a comment, synonymous with REM in DOS.

Add the following to the top of the mrtg.cfg file:

 WorkDir: D:\InetPub\wwwroot\MRTG

This is where the web pages are created, usually a web root.

 ######################################################################
 # Description: LCP SUWGB
 # Contact: Administrator
 # System Name: LC-Bridge
 # Location: Here
 #.....................................................................

TargetDevice's IP Address:Interface Number:Community:IP Address

 Target[10.10.10.1.1]: 1:public@10.10.10.1

This is the interface speed (Default is 10 megabits; for 100Mbit devices use 12500000 and so on...)

 MaxBytes[10.10.10.1.1]: 1250000
 Title[10.10.10.1.1]: LC-Bridge (sample.device): ether0

This section determines how the web page headers will look

 PageTop[10.10.10.1.1]: <H1>Traffic Analysis for ether0</H1>
  <TABLE>
  <TR><TD>System:</TD><TD>LC-Bridge inAndover</TD></TR>
  <TR><TD>Maintainer:</TD><TD>Administrator</TD></TR>
  <TR><TD>Interface:</TD><TD>ether0(1)</TD></TR>
  <TR><TD>IP:</TD><TD>sample.device(10.10.10.1)</TD></TR>
  <TR><TD>Max Speed:</TD>
  <TD>1250.0 kBytes/s (ethernetCsmacd)</TD></TR>
  </TABLE>
  Target[10.10.10.1.2]: 2:public@10.10.10.1
  MaxBytes[10.10.10.1.2]: 1250000
  Title[10.10.10.1.2]: LC-Bridge (): ulink0
  PageTop[10.10.10.1.2]: <H1>Traffic Analysis for ulink0</H1>
   <TABLE>
   <TR><TD>System:</TD><TD>LC-Bridge inAndover</TD></TR>
   <TR><TD>Maintainer:</TD><TD>Administrator</TD></TR>
   <TR><TD>Interface:</TD><TD>ulink0(2)</TD></TR>
   <TR><TD>IP:</TD><TD>()</TD></TR>
   <TR><TD>Max Speed:</TD>
   <TD>1250.0 kBytes/s (ethernetCsmacd)</TD></TR>
   </TABLE>
  #---------------------------------------------------------------

And that is a very basic mrtg config file. You can run this and see your results by going into the c:\mrtg-2.9.17\bin directory and typing:

 perl mrtg mrtg.cfg

It is normal to get errors for the first two times you run this command. The errors will alert you about the fact that there have not been any log files in existnace before.

If you take a look at those web pages they are not real exciting, yet. You need to have the mrtg files run every five minutes to produce the desired results. Just run it again after a few minutes. You should now be able to see the first lines in your graphs.


MAKE MRTG RUN ALL THE TIME

Starting mrtg by hand every time you want to run it is not going to make you happy I guess.

There is a special option you can set in the mrtg configuration file so so that mrtg will not terminate after it was started. Instead it will wait for 5 minutes and then run again.

Add the option

 RunAsDaemon: yes

to your mrtg.cfg file and start it with:

 start /Dc:\mrtg-2.9.17\bin wperl mrtg --logging=eventlog mrtg.cfg

If you use wperl instead of perl, no console window will show. MRTG is now running in the background. If it runns into problems it will tell you so over the EventLog. To stop MRTG, open the taskmanager and terminate the wperl.exe process. If mrtg has anything to tell you these messages can be found in the eventlog.

If you put a shortcut with

 Target:    wperl mrtg --logging=eventlog mrtg.cfg
 Start in:  c:\mrtg-2.9.17\bin

into your startup folder, mrtg will now start whever you login to your NT box.

If you do not want to log into your NT box just to start mrtg. Have a look at http://www.firedaemon.com/mrtg-howto.html which describes a free tool to start any program as a Service. The pages gives specific instructions for mrtg users.


EXAMPLE

Now lets look at a config file to monitor what we wanted to on our mythical Cisco Cat 5000 -- utilization on ports 3, 5, 10, and 24, and the CPU Load, which will show us nonstandard mrtg configurations as well as more options..

 WorkDir: D:\InetPub\wwwroot\MRTG
 ######################################################################
 # Description: LCP SUWGB
 # Contact: Administrator
 # System Name: LC-Bridge
 # Location: Here
 #.....................................................................
 Target[10.10.10.1.1]: 3:public@10.10.10.1
 MaxBytes[10.10.10.1.1]: 1250000
 Title[10.10.10.1.1]: LC-Bridge (sample-device): ether0
 PageTop[10.10.10.1.1]: <H1>Traffic Analysis for ether0</H1>
  <TABLE>
 <TR><TD>System:</TD><TD>LC-Bridge inAndover</TD></TR>
 <TR><TD>Maintainer:</TD><TD>Administrator</TD></TR>
 <TR><TD>Interface:</TD><TD>ether0(3)</TD></TR>
 <TR><TD>IP:</TD><TD>sample-device(10.10.10.1)</TD></TR>
 <TR><TD>Max Speed:</TD>
 <TD>1250.0 kBytes/s (ethernetCsmacd)</TD></TR>
 </TABLE>
 #---------------------------------------------------------------
 Target[10.10.10.1.2]: 5:public@10.10.10.1
 MaxBytes[10.10.10.1.2]: 1250000
 Title[10.10.10.1.2]: LC-Bridge (): ulink0
 PageTop[10.10.10.1.2]: <H1>Traffic Analysis for ulink0</H1>
  <TABLE>
  <TR><TD>System:</TD><TD>LC-Bridge inAndover</TD></TR>
  <TR><TD>Maintainer:</TD><TD>Administrator</TD></TR>
  <TR><TD>Interface:</TD><TD>ulink0(5)</TD></TR>
  <TR><TD>IP:</TD><TD>()</TD></TR>
  <TR><TD>Max Speed:</TD>
  <TD>1250.0 kBytes/s (ethernetCsmacd)</TD></TR>
  </TABLE>
 #---------------------------------------------------------------
 Target[10.10.10.1.1]: 10:public@10.10.10.1
 MaxBytes[10.10.10.1.1]: 1250000
 Title[10.10.10.1.1]: LC-Bridge (sample-device): ether0
 PageTop[10.10.10.1.1]: <H1>Traffic Analysis for ether0</H1>
  <TABLE>
  <TR><TD>System:</TD><TD>LC-Bridge inAndover</TD></TR>
  <TR><TD>Maintainer:</TD><TD>Administrator</TD></TR>
  <TR><TD>Interface:</TD><TD>ether0(10)</TD></TR>
  <TR><TD>IP:</TD><TD>sample-device(10.10.10.1)</TD></TR>
  <TR><TD>Max Speed:</TD>
  <TD>1250.0 kBytes/s (ethernetCsmacd)</TD></TR>
  </TABLE>
 #---------------------------------------------------------------
 Target[10.10.10.1.2]: 24:public@10.10.10.1
 MaxBytes[10.10.10.1.2]: 1250000
 Title[10.10.10.1.2]: LC-Bridge (): ulink0
 PageTop[10.10.10.1.2]: <H1>Traffic Analysis for ulink0</H1>
  <TABLE>
  <TR><TD>System:</TD><TD>LC-Bridge inAndover</TD></TR>
  <TR><TD>Maintainer:</TD><TD>Administrator</TD></TR>
  <TR><TD>Interface:</TD><TD>ulink0(24)</TD></TR>
  <TR><TD>IP:</TD><TD>()</TD></TR>
  <TR><TD>Max Speed:</TD>
  <TD>1250.0 kBytes/s (ethernetCsmacd)</TD></TR>
  </TABLE>
 #---------------------------------------------------------------
 # Router CPU load %
 Target[cpu.1]:1.3.6.1.4.1.9.2.1.58.0&1.3.6.1.4.1.9.2.1.58.0:public@10.10.10.1
 RouterUptime[cpu.1]: public@10.10.10.1
 MaxBytes[cpu.1]: 100
 Title[cpu.1]: CPU LOAD
 PageTop[cpu.1]: <H1>CPU Load %</H1>
 Unscaled[cpu.1]: ymwd
 ShortLegend[cpu.1]: %
 XSize[cpu.1]: 380
 YSize[cpu.1]: 100
 YLegend[cpu.1]: CPU Utilization
 Legend1[cpu.1]: CPU Utilization in % (Load)
 Legend2[cpu.1]: CPU Utilization in % (Load)
 Legend3[cpu.1]:
 Legend4[cpu.1]:
 LegendI[cpu.1]:
 LegendO[cpu.1]: &nbsp;Usage
 Options[cpu.1]: gauge

This is a nice example of how to monitor any SNMP device if you know what OID you want to use. Once again, For an explanation of the more advance features of mrtg, please see Tobias's documentation.


AUTHORS

David S. Divins <ddivins@moon.jic.com>, Steve Pierce <MRTG@HDL.com>, Tobi Oeitker <oetiker@ee.ethz.ch>