Scripting help...monitori ng disk space

[asmpic]

Hi,
I stumbled upon a script on the internet that monitors disk space and if it reaches a certain percentage being full it send an email. However, the problem I encountered was that one of the file system names are longer than usual so it expends over to the other columns and causes the other data to be put onto a second line. So the script no longer works when this is the case. does anyone know how to fix the script to take this into account?

This is what I get when I run df-H. The first line goes onto a second line because of the file name so the script return incorrect information when I ask to print columns 5 and 1.

Filesystem             Size   Used  Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                        77G   8.2G    65G  12% /
/dev/sda1              104M    14M    86M  14% /boot
none                   525M      0   525M   0% /dev/shm
 

This is the script I'm using.

#!/bin/sh
df -H | grep -vE '^Filesystem|none|cdrom' | awk '{ print $5 " " $1 }' | while read output;
do
   #echo $output
   usep=$(echo $output | awk '{ print $1}' | cut -d'%' -f1  )
   partition=$(echo $output | awk '{ print $2 }' )
   if [ $usep -ge 10 ]; then
   echo "Running out of space \"$partition ($usep%)\" on $(hostname)" | mail -s "Alert: Almost out of disk space $usep" [e-mail]
fi
done

[kvkrishna]

Hi,
I use the below to monitor the disk space....

df -kt | grep "%" | grep "^/dev" >/tmp/file_sys.txt
chmod 755 /tmp/file_sys.txt
Flag=0
while read LINE
do
        pcnt=`echo $LINE | cut -d"%" -f1 | awk '{print $NF}'`
        f_sys=`echo $LINE | cut -d"%" -f2 | awk '{print $NF}'`
        if [ $Flag -eq 0 ]
        then
        if [ $pcnt -ge 80 ] && [ $pcnt -le 100 ]
        then
                echo $dashes >/tmp/`hostname`_alert.txt
                echo "The below file systems need house keeping.....!" >>/tmp/`hostname`_alert.txt
                echo $dashes >>/tmp/`hostname`_alert.txt
                #echo "File System $f_sys is $pcnt%"
                Flag=1
        else
                Flag=0
        fi
        fi
        if [[ $Flag -eq 1 ]]
        then
                if [ $pcnt -ge 80 ] && [ $pcnt -le 100 ]
                then
                        echo "File System $f_sys is $pcnt%" >>/tmp/`hostname`_alert.txt
                fi
        fi
done </tmp/file_sys.txt
if [ -f /tmp/`hostname`_alert.txt ]
then
        if [ -s /tmp/`hostname`_alert.txt ]
        then
                #cat /tmp/`hostname`_alert.txt
                mail_subject="ITRM:Alert from vista70"
                mail_file=/tmp/`hostname`_alert.txt
                mail_recp="[e-mail] [e-mail] [e-mail]" 
                mailx -s "$mail_subject" $mail_recp < $mail_file
        fi
fi
 

Give me output as:

+-----------------------------------------------------------------------------+
The below file systems need house keeping.....!
+-----------------------------------------------------------------------------+
File System /vandata02 is 81%
File System /datafi01 is 84%
File System /data4 is 93%
File System /vanfi is 82%
File System /vandev is 91%
File System /VANPOC is 91%
File System /datauat01 is 88%
File System /datauat02 is 87%
File System /datahotfix01 is 96%
File System /datavanpc01 is 82%

Hi,
I stumbled upon a script on the internet that monitors disk space and if it reaches a certain percentage being full it send an email. However, the problem I encountered was that one of the file system names are longer than usual so it expends over to the other columns and causes the other data to be put onto a second line. So the script no longer works when this is the case. does anyone know how to fix the script to take this into account?

This is what I get when I run df-H. The first line goes onto a second line because of the file name so the script return incorrect information when I ask to print columns 5 and 1.

Filesystem             Size   Used  Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                        77G   8.2G    65G  12% /
/dev/sda1              104M    14M    86M  14% /boot
none                   525M      0   525M   0% /dev/shm
 

This is the script I'm using.

#!/bin/sh
df -H | grep -vE '^Filesystem|none|cdrom' | awk '{ print $5 " " $1 }' | while read output;
do
   #echo $output
   usep=$(echo $output | awk '{ print $1}' | cut -d'%' -f1  )
   partition=$(echo $output | awk '{ print $2 }' )
   if [ $usep -ge 10 ]; then
   echo "Running out of space \"$partition ($usep%)\" on $(hostname)" | mail -s "Alert: Almost out of disk space $usep" [e-mail]
fi
done

[optikknight]

I've occasionally had the same problem... (And yes, I know I'm waking a dead thread. I figured I'd share my solution anyway.)

Here's my script for the same purpose:
df.cron

It assumes all lines not containing a '%' are wrapped, and pulls the next line. Example output:

Nearly full filesystems:
Filesystem               Size  Used  Avail  Use%  Mounted-on
/dev/mapper/System-home  100G  69G   26G    73%   /home
/dev/hda6                38G   34G   2.3G   94%   /mnt/unsafe

(The write_preformat ted call is a routine from the included lib, in the linked script.)

The more interesting parts, for text processing:

df -h > $filesystems
sed '1d; /%/! { N; s/[\n\t]/ /; }' $filesystems | grep -E '([7-9][0-9]|[0-9]{3})%' > $filesystems.2