Memory check plugin for Nagios

Mike Adolphs — Thu, 26 Feb 2009 10:50:30 +0000

As I’m currently reworking our large, distributed network monitoring solution, I’m also integrating own plugins for various purposes. The beginning is made by a check plugin for memory consumption on Linux hosts which checks /proc/meminfo directly instead of other checks which usually use frontends like free or top. Besides that, there’s really no magic (and parsing /proc/meminfo isn’t magic as well). You’re able to switch between kilobyte and percentage output, defining warning and critical levels is also possible.
It is sh-compliant, released under the terms of the GPLv2 and as every well written shell script there’s no need to provide any argument at all. For full help use -h or –help.

Give it a try if you like and let me know if it doesn’t work as intended.
Please copy the content below. The script will be reworked and uploaded to Nagios and MonitoringExchange after my vacation at the beginning of July.

#!/bin/sh                                           
 
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License as published by
#   the Free Software Foundation; version 2 of the License only.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
ST_OK=0
ST_WR=1
ST_CR=2
ST_UK=3
wcdiff=0
wclvls=0
 
PROGNAME=`basename $0`
VERSION="Version 1.0,"
AUTHOR="2009, Mike Adolphs (http://www.matejunkie.com/)"
 
print_version() {
    echo "$VERSION $AUTHOR"
}                          
 
print_help() {
    print_version $PROGNAME $VERSION
    echo ""
    echo "$PROGNAME is a Nagios plugin to check the current memory usage via"
    echo "/proc/meminfo. It should work on almost every system where you're"
    echo "able to cat /proc/meminfo. The script itself is written sh-compliant"
    echo "and free software under the terms of the GPLv2."
    echo ""
    echo "$PROGNAME -s/--style kb/perc [-w/--warning] [-c/--critical]"
    echo "Options:"
    echo "  -s/--style)"
    echo "     You're able to set the output to kilobyte (kb) or percentage"
    echo "     (perc) values. Please note that other values will result in"
    echo "     termination. Default is: kb"
    echo "  -w/--warning)"
    echo "     Warning level in kb or percent for the amount of used memory."
    echo "     Note that when the style is set to percentage, only values"
    echo "     betweeen 0 and 100 are allowed. Default is: Off."
    echo "  -c/--critical)"
    echo "     Critical level in kb or percent for the amount of used memory."
    echo "     Note that when the style is set to percentage, only values"
    echo "     between 0 and 100 are allowed. Default is: Off."
    exit $ST_UK
}                                                                              
 
val_wcdiff() {
    if [ ${lv_wr} -gt ${lv_cr} ]
    then
        wcdiff=1
    fi
}                               
 
val_wclvls() {
    if [ $style = "perc" ]
    then
        if [ "$lv_wr" -lt 0 -o "$lv_wr" -gt 100 -o "$lv_cr" -lt 0 -o "$lv_cr" -gt 100 ]
        then
            wclvls=1
            val_wcdiff
        fi
    fi
}                                                                                      
 
style="kb"
 
while test -n "$1"; do
    case "$1" in
        --help|-h)
            print_help
            exit $ST_UK
            ;;
        --version|-v)
            print_version $PROGNAME $VERSION
            exit $ST_UK
            ;;
        --style|-s)
            style=$2
            shift
            ;;
        --warning|-w)
            lv_wr=$2
            shift
            ;;
        --critical|-c)
            lv_cr=$2
            shift
            ;;
        *)
            echo "Unknown argument: $1"
            print_help
            exit $ST_UK
            ;;
    esac
    shift
done                                        
 
if [ ! -z "$lv_wr" -a ! -z "$lv_cr" ]
then
    val_wcdiff
    val_wclvls
fi                                   
 
if [ $wcdiff = 1 ]
then
    echo "Please adjust levels. The critical level must be higher than the warning level!"
    if [ $wclvls = 1 ]
    then
        echo "Warning and critical level values must be between 0 and 100."
    fi
    exit $ST_UK
fi                                                                                        
 
if [ $wclvls = 1 ]
then
    echo "Warning and critical level values must be between 0 and 100."
    exit $ST_UK
fi                                                                     
 
MEM_TOTAL=`grep "^MemTotal" /proc/meminfo|awk '{print $2}'`
TMP_MEM_FREE=`grep "^MemFree" /proc/meminfo|awk '{print $2}'`
TMP_MEM_USED=`expr $MEM_TOTAL - $TMP_MEM_FREE`
BUFFERS=`grep "^Buffers" /proc/meminfo|awk '{print $2}'`
CACHED=`grep "^Cached" /proc/meminfo|awk '{print $2}'`       
 
P_MEM_FREE=`echo "scale=2; $TMP_MEM_FREE / $MEM_TOTAL * 100" | bc -l | sed 's/.[0-9][0-9]//'`
P_MEM_USED=`echo "scale=0; 100 - $P_MEM_FREE" | bc -l`                                       
 
if [ $style = "kb" ]
then
    if [ ! -z "$lv_wr" -a ! -z "$lv_cr" ]
    then
        if [ ${TMP_MEM_USED} -ge ${lv_wr} -a ${TMP_MEM_USED} -lt ${lv_cr} ]
        then
            echo "WARNING - Total: $MEM_TOTAL kb, Used: $TMP_MEM_USED kb, Free: $TMP_MEM_FREE kb | 'mem_total'=$MEM_TOTAL 'mem_used'=$TMP_MEM_USED;$lv_wr;$lv_cr 'mem_free'=$TMP_MEM_FREE"
            exit $ST_WR
        elif [ ${TMP_MEM_USED} -ge ${lv_cr} ]
        then
            echo "CRITICAL - Total: $MEM_TOTAL kb, Used: $TMP_MEM_USED kb, Free: $TMP_MEM_FREE kb | 'mem_total'=$MEM_TOTAL 'mem_used'=$TMP_MEM_USED;$lv_wr;$lv_cr 'mem_free'=$TMP_MEM_FREE"
            exit $ST_CR
        else
            echo "OK - Total: $MEM_TOTAL kb, Used: $TMP_MEM_USED kb, Free: $TMP_MEM_FREE kb | 'mem_total'=$MEM_TOTAL 'mem_used'=$TMP_MEM_USED;$lv_wr;$lv_cr 'mem_free'=$TMP_MEM_FREE"
            exit $ST_OK
        fi
    else
        echo "OK - Total: $MEM_TOTAL kb, Used: $TMP_MEM_USED kb, Free: $TMP_MEM_FREE kb | 'mem_total'=$MEM_TOTAL 'mem_used'=$TMP_MEM_USED 'mem_free'=$TMP_MEM_FREE"
        exit $ST_OK
    fi
elif [ $style = "perc" ]
then
    if [ ! -z "$lv_wr" -a ! -z "$lv_cr" ]
    then
        if [ ${P_MEM_USED} -ge ${lv_wr} -a ${P_MEM_USED} -lt ${lv_cr} ]
        then
            echo "WARNING - Used: $P_MEM_USED%, Free: $P_MEM_FREE% | 'mem_used'=$P_MEM_USED;$lv_wr;$lv_cr 'mem_free'=$P_MEM_FREE"
            exit $ST_WR
        elif [ ${P_MEM_USED} -ge ${lv_cr} ]
        then
            echo "CRITICAL - Used: $P_MEM_USED%, Free: $P_MEM_FREE% | 'mem_used'=$P_MEM_USED;$lv_wr;$lv_cr 'mem_free'=$P_MEM_FREE"
            exit $ST_CR
        else
            echo "OK - Used: $P_MEM_USED%, Free: $P_MEM_FREE% | 'mem_used'=$P_MEM_USED;$lv_wr;$lv_cr 'mem_free'=$P_MEM_FREE"
            exit $ST_OK
        fi
    else
            echo "OK - Used: $P_MEM_USED%, Free: $P_MEM_FREE% | 'mem_used'=$P_MEM_USED 'mem_free'=$P_MEM_FREE"
            exit $ST_OK
    fi
else
    echo "Style (-s) must be either kb (kilobyte) or perc (for percent). kb is"
    echo "being used as the default behavior when you don't provide the -s option."
    echo ""
    echo "For more information try -h or --help!"
    exit $ST_UK
fi

Magic SysRq – REISUB

Mike Adolphs — Mon, 09 Feb 2009 03:40:35 +0000

Ever encoutered a frozen X-Server without the possibility to switch to another console or a rare, but nasty kernel panic? Probably and I guess the majority of the people hit the power button in those cases risking data loss.

This is not necessary. There are other ways to restart a frozen Linux system!

The kernel has a feature called “Magic System Request Keys” – if it was compiled with the CONFIG_MAGIC_SYSRQ option though. It is available via ALT+PRINT on most systems and provides various low level commands to safely turn of or reboot your computer in case of a failure.

If you want to know whether this feature is available on your machine, ask /proc for it. If it returns a 1 Magic SysRq is active, a 0 indicates that it’s not.

user@server:~$ cat /proc/sys/kernel/sysrq
1

Hold down ALT+PRINT and type slowly REISUB to restart your system. Hold down ALT+PRINT and type REISUO to turn it off.

A short note on what REISUB or REISUO does:

R – Switches the keyboard from raw to XLATE mode (detatching from the X-Server)
E – Sending a SIGTERM to all processes except INIT
I – Sending a SIGKILL to all processes except INIT (and that’s why to type REISUB slowly)
S – Syncs all data in the cache to mounted partitions
U – Unmounts all mounted partitions and remounts them read-only
B – Reboot the system immediately without unmounting or syncing mounted partitions
or
O – Shut down the system immediately without unmounting or syncing mounted partitions

Be aware of the fact that the feature is currently broken in Ubuntu 8.10 because of a problem regarding kbd, evdev and 8.10′s xserver version. It only works from a text console, therefore not on the X-Server.
It is possible to reconfigure the xorg.conf from kbd to evdev like in the following, but no guarantees:

Section "InputDevice"
	Identifier	"Generic Keyboard"
	Driver		"kbd"
	Option		"XkbRules"	"xorg"
	Option		"XkbModel"	"pc105"
	Option		"XkbLayout"	"de"
EndSection

Section "ServerLayout"
        ...
	InputDevice	"Generic Keyboard"
	Option		"AutoAddDevices"	"false"
EndSection

It’ll be fixed in 9.04 coming this April as stated in the bug description on Launchpad.

One last hint: If your machine doesn’t have a PRINT/SysRq key and you’re still able to access the console, yu may also send the commands via echo to the sysrq-trigger as seen in the example below:

user@server:~$ echo r | sudo tee /proc/sysrq-trigger

Matejunkie » kernel

Memory check plugin for Nagios

Magic SysRq – REISUB