mrtg-2.12.2-2 (Monitoring)

Updated on 18 Jan, 2006 by SuSE 10.0

SNMP

Make sure net-snmp is installed, configured, and running.

On Linux, net-snmp-5.2.1-10.i586.rpm is installed, and configuration would look like this (etc/snmp.conf).

#
# First, map the community name "public" into a "security name"
#	sec.name source	community
com2sec	public	default	public

# Second, map the security name into a group name.
#	groupName	securityModel	securityName
group	public		v1		public
group	public		v2c		public
group	public		usm		public

# Third, create a view for us to let the group have rights to:
#	name	incl/excl	subtree	mask(optional)
view	all	included	.1

# Finally, grant the group read-only access to the systemview view.
#	group	context	sec.model	  sec.level prefix	read	write	notif
access	public	""	any		noauth	exact	all	all	all

# System contact information
#
syslocation	xxxxxxxx
syscontact	foo@host

On Windows 2000, XP, 2003, SNMP Service should be installed, and it would look like this,

On CISCO products, SNMP configuration would like like this,

gw(config)#snmp-server community public RO

Make sure, snmp is running and accessible. If necessary, open 161 tcp, udp ports on the target device.

# snmpwalk -v 1 192.168.1.100 -c public | less
SNMPv2-MIB::sysDescr.0 = STRING: Hardware: x86 Family 6 Model 9 Stepping 5 AT/AT CO
MPATIBLE - Software: Windows 2000 Version 5.1 (Build 2600 Uniprocessor Free)
SNMPv2-MIB::sysObjectID.0 = OID: SNMPv2-SMI::enterprises.311.1.1.3.1.1
DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (80732) 0:1
SNMPv2-MIB::sysContact.0 = STRING:
SNMPv2-MIB::sysName.0 = STRING: TS-NEC
SNMPv2-MIB::sysLocation.0 = STRING:
SNMPv2-MIB::sysServices.0 = INTEGER: 76
IF-MIB::ifNumber.0 = INTEGER: 4
....

Installation

Install from SuSE installation source

# rpm -ihv mrtg-2.12.2-2.i586.rpm

Configuration

The next step is now to configure mrtg for monitoring an network device.

Make sure net-snmp is installed, configured, and running.

Make mrtg directory

# mkdir /srv/www/htdocs/mrtg

Make configuration file and run it (/etc/mrtg/makecfg.sh)

#!/bin/bash

# host list
HOSTLIST="
192.168.0.1
192.168.0.2
"

# Make configuration file
for HOST in $HOSTLIST
do
        /usr/bin/cfgmaker \
                --ifref=eth \
                --global 'WorkDir: /srv/www/htdocs/mrtg' \
                --global 'Options[_]: bits,growright' \
                --output /etc/mrtg/$HOST.cfg \
                public@$HOST
done

This example monitor two devices (192.168.0.1, 192.168.0.2) and makes two configuration files (192.168.0.1.cfg and 192.168.0.2.cfg)

Run

Make an executing file (/etc/mrtg/mrtg.cron) to run

#!/bin/bash

DIRCFG=/etc/mrtg
DIRHTML=/srv/www/htdocs/mrtg
LISTCFG="
192.168.0.1.cfg
192.168.0.2.cfg
"

for CFG in $LISTCFG
do
        env LANG=C /usr/bin/mrtg $DIRCFG/$CFG
done

# Make index file
cd $DIRCFG
/usr/bin/indexmaker --columns=1 $LISTCFG > $DIRHTML/index.html

To run every 5 minutes, add one line in /etc/crontab

...
0-59/5 * * * * root /etc/mrtg/mrtg.cron >/dev/null 2>&1

In the default configuration, only the localhost can view the analysis from http://3/mrtg. However, if you want to allow other computer to view it, edit /etc/httpd/conf.d/mrtg.conf as below,

    Order deny,allow
#    Deny from all
    Allow from localhost
    # Allow from .example.com

Monitor CPU

Add following lines to the configuration file.

# This section creates CPU Load monitoring
### CPU Load Average ###
Target[cpu_hostname]: .1.3.6.1.4.1.2021.10.1.5.1&.1.3.6.1.4.1.2021.10.1.5.2:public@localhost
MaxBytes[cpu_hostname]: 100
Unscaled[cpu_hostname]: dwmy
Options[cpu_hostname]: gauge, absolute, growright, noinfo, nopercent
YLegend[cpu_hostname]: CPU Load(%)
ShortLegend[cpu_hostname]: (%)
LegendI[cpu_hostname]: 1min ave
LegendO[cpu_hostname]: 5min ave
Legend1[cpu_hostname]: 1min ave(%)
Legend2[cpu_hostname]: 5min ave(%)
Title[cpu_hostname]: CPU Load ($HOST)
PageTop[cpu_hostname]: <H1>CPU Load Average -- suse10</H1>

Monitor Disk usage without snmp

Create a script at /usr/local/sbin/df.pl

#!/usr/bin/perl

# It assumes that the command output(df -kl) looks like this:
# Filesystem           1k-blocks      Used Available Use% Mounted on
# /dev/md0              95645100  30401312  64272080  33% /
# /dev/hde1                14119      1159     12231   9% /boot
#
# In which case, this script returns :
# 95659219
# 30402503
# when run.

foreach $filesystem (`df -kl | grep -v "Filesystem"`)
{
  @df = split(/\s+/,$filesystem);
  $total += $df[1];
  $usage += $df[2];
}
print "$usage\n";
print "$usage\n";
print "$total\n";

Give permission.

# chmod 755 df.pl

create xinetd file at /etc/xinet.d/df

service df
{
        socket_type     = stream
        protocol        = tcp
        wait            = no
        user            = nobody
        bind            = 127.0.0.1
        server          = /usr/local/sbin/df.pl
}

Add a line to /etc/services

df              9047/tcp                        # df

Restart xinet.d

# /etc/init.d/xinetd restart

Test the configuration

# /usr/bin/netcat localhost df
3479420
3479420
9931728

Add followings to the MRTG configuration file

# Disk usage monitoring
Target[disk_hostname]: `/usr/bin/netcat localhost df`
Title[disk_hostname]: Disk Usage -- suse10
Unscaled[disk_hostname]: dwym
MaxBytes[disk_hostname]: 9931728
PageTop[disk_hostname]: <H1>Disk Usage -- suse10</H1>
kmg[disk_hostname]: KB,MB,GB
LegendI[disk_hostname]: Used Disk Space
LegendO[disk_hostname]: Total Disk Space
Legend1[disk_hostname]: Used Disk Space
Legend2[disk_hostname]: Total Disk Space
YLegend[disk_hostname]:  Megabytes
ShortLegend[disk_hostname]:  
Options[disk_hostname]: gauge, growright, noinfo

Monitor Memory usage without snmp

Make sure you have sysstat-5.0.6-6.i586.rpm installed

Create a script at /usr/local/sbin/memory.pl

#!/usr/bin/perl

foreach $_ (`/usr/bin/sar -r 1 2 | grep "Average"`)
{
  ($average, $kbmemfree, $kbmemused, $memused, $kbmemshrd, $kbbuffers, $kbcached, $kbswpfree, $kbswpused, $swpused) = split(/\s+/);
  print $kbmemused - $kbbuffers . "\n";
  print $kbmemused . "\n";
  print $kbmemused + $kbmemfree . "\n";
}

Give permission.

# chmod 755 memory.pl

create xinetd file at /etc/xinet.d/memory

service memory
{
        socket_type     = stream
        protocol        = tcp
        wait            = no
        user            = nobody
        bind            = 127.0.0.1
        server          = /usr/local/sbin/memory.pl
}

Add a line to /etc/services

memory          9057/tcp                        # memory

Restart xinet.d

# /etc/init.d/xinetd restart

Test the configuration

# /usr/bin/netcat localhost memory
203578
308974
312960

Add followings to the MRTG configuration file

# Memory usage monitoring
Target[memory_hostname]: `/usr/bin/netcat localhost memory`
MaxBytes[memory_hostname]:  312960
Unscaled[memory_hostname]:  dwym
YLegend[memory_hostname]:  Memory Usage
ShortLegend[memory_hostname]:  
kmg[memory_hostname]:  kB,MB
kilo[memory_hostname]:  1024
LegendI[memory_hostname]:  App Data
LegendO[memory_hostname]:  Memory Used
Legend1[memory_hostname]:  App Data
Legend2[memory_hostname]:  Memory Used
Options[memory_hostname]: gauge, growright, noinfo
Title[memory_hostname]: Memory Usage -- suse10
PageTop[memory_hostname]: <H1>Memory Usage -- suse10</H1>

Monitor Swap usage without snmp

Create a script at /usr/local/sbin/swap.pl

#!/usr/bin/perl

foreach $_ (`/usr/bin/sar -r 1 2 | grep "Average"`)
{
  ($average, $kbmemfree, $kbmemused, $memused, $kbbuffers, $kbcached, $kbswpfree, $kbswpused, $swpused, $kbswpcad) = split(/\s+/);
  print $kbswpcad . "\n";
  print $kbswpused . "\n";
  print $kbswpused + $kbswpfree . "\n";
}

Give permission.

# chmod 755 swap.pl

create xinetd file at /etc/xinet.d/swap

service swap
{
        socket_type     = stream
        protocol        = tcp
        wait            = no
        user            = nobody
        bind            = 127.0.0.1
        server          = /usr/local/sbin/swap.pl
}

Add a line to /etc/services

swap            9058/tcp                        # swap

Restart xinet.d

# /etc/init.d/xinetd restart

Test the configuration

# /usr/bin/netcat localhost swap
40
3104
706820

Add followings to the MRTG configuration file

# Swap usage monitoring
Target[swap_hostname]: `/usr/bin/netcat localhost swap`
MaxBytes[swap_hostname]: 706820
Unscaled[swap_hostname]: dwym
YLegend[swap_hostname]: Swap In Use
kmg[swap_hostname]:  kB,MB
kilo[swap_hostname]:  1024
LegendI[swap_hostname]:  Cached Swap
LegendO[swap_hostname]:  Swap Used
Legend1[swap_hostname]:  Cached Swap
Legend2[swap_hostname]:  Swap Used
Options[swap_hostname]: gauge, growright, noinfo
ShortLegend[swap_hostname]: 
Title[swap_hostname]: Swap Usage -- suse10
PageTop[swap_hostname]: <H1>Swap Usage -- suse10</H1>

Monitor Network Availability (ping)

Install mtr package which comes with SuSE. mtr is a network diagnostic tool.

# rpm -ihv mtr-0.69-4.i586.rpm

Create a script as, /usr/local/sbin/ping.pl

#!/usr/bin/perl
# Usage: perl.pl host

@result = split(/\s+/,`/usr/sbin/mtr -r $ARGV[0] | tail -n 1`);
if ($result[7] != 0) {
        print $result[7] / $result[8] * 100 . "\n";
        print 100-$result[3] . "\n";
} else {
        print "0\n";
        print "0\n";
}

Change the permission

# chmod 755 ping.pl

Add following into the mrtg configuration file. This monitor www.hoge.com

# Network Availability
Target[ava_hoge]: `/usr/local/sbin/ping.pl www.hoge.com`
MaxBytes[ava_hoge]: 100
Unscaled[ava_hoge]: dwym
YLegend[ava_hoge]: Availability
ShortLegend[ava_hoge]: (%)
LegendI[ava_hoge]:  Stability (best/worst)
LegendO[ava_hoge]:  Availability
Legend1[ava_hoge]:  Stability (best/worst)
Legend2[ava_hoge]:  Availability
Options[ava_hoge]: gauge, absolute, growright, noinfo, nopercent
Title[ava_hoge]: Availability -- www.hoge.com
PageTop[ava_hoge]: <H1>Availability -- www.hoge.com</H1>

This configuration checks connectivity for 10 seconds (ten pings on each second) on every five minues.

To monitor multiple destinations, repeat the configuration, but change the bold place.

Back