Disk space alert

rbalaa · 09-02-2011, 12:36 PM

Hi All,

I have a few servers running and would like to set an alert to let me know when the disk space is close to be full.

Can anyone suggest a solution ? Thanks.

thesnow · 09-02-2011, 12:46 PM

You could set up a job in cron on each server to check periodically and send you an email.

Or, you could set up a central host that ssh's to a list of server and checks, then notifies you of any that are nearing your threshold.

Or, you could set up something like Nagios http://www.nagios.org/.

ncsuapex · 09-02-2011, 01:52 PM

One small problem with doing

Quote:

You could set up a job in cron on each server to check periodically and send you an email.

is if the disk space fills before your cron job runs you wont get an email because it wont have the tmp space to compose the mail. Nagios is good to monitor disk space.

rbalaa · 09-02-2011, 04:22 PM

I installed Nagios and looking at it now. But I'm really confused. Looking for a help guide of some sort just to get an overview

Got the install instructions from here if any is interested: http://nagios.sourceforge.net/docs/3...rt-fedora.html

frankbell · 09-02-2011, 07:42 PM

There's lots of docs at the Nagios site.

http://www.nagios.org/documentation

amachina · 09-02-2011, 08:39 PM

Code:

#!/bin/bash
#
# Name: chk_fs.sh
# Schedule: Every 20 minutes
# Function: Check all filesystems except /uxx; alert if used % exceeds PCT

PCT=85
typeset -a ARR
PVAL="/tmp/chkfs1.txt"
CVAL="/tmp/chkfs2.txt"
FSTYP=ext4
RECIPIENTS="user@host"

# gather current data
STRING=$(df -hPT | \
   grep ${FSTYP} | \
   awk '{print $2,$6,$7}' | \
   while read FSTYPE FSUSED FSNAME
   do
      if [[ "${FSNAME}" =~ "/u[0-9][0-9]" ]]
      then
         :
      else
         PCTUSD=$(echo ${FSUSED} | tr -d "%")
         if [[ ${PCTUSD} -ge ${PCT} ]]
         then
            echo "${FSNAME}:${FSTYPE}:${PCTUSD}"
         fi
      fi
   done)

# put current data in an array
((i=0))
for x in $(echo ${STRING})
do
   ARR[$i]=${x}
   ((i=i+1))
done
LENG=${#ARR[*]}

# check prior data against array
> ${CVAL}
if [[ ${LENG} -gt 0 ]]
then
   while read FSN USD REST
   do
      ((i=0))
      while [[ ${i} -lt ${LENG} ]]
      do
         CURFSN=$(echo ${ARR[$i]} | cut --delimiter=: --fields=1)
         CURUSD=$(echo ${ARR[$i]} | cut --delimiter=: --fields=3)
         if [[ "${FSN}" == "${CURFSN}" ]]
         then
            if [[ ${CURUSD} -ge ${USD} ]]
            then
               if [[ ${CURUSD} -gt ${USD} ]]
               then
                  echo -e "${FSN}\t\t${CURUSD}\t\t+" >> ${CVAL}
               else
                  echo -e "${FSN}\t\t${CURUSD}\t\t=" >> ${CVAL}
               fi
               ARR[$i]=0
            fi
         fi
         ((i=i+1))
      done
   done < ${PVAL}
fi

# check for new entries in array
((i=0))
while [[ ${i} -lt ${LENG} ]]
do
   if [[ "${ARR[$i]}" != "0" ]]
   then
      CURFSN=$(echo ${ARR[$i]} | cut --delimiter=: --fields=1)
      CURUSD=$(echo ${ARR[$i]} | cut --delimiter=: --fields=3)
      echo -e "${CURFSN}\t\t${CURUSD}\t\t!" >> ${CVAL}
   fi
   ((i=i+1))
done

cp ${CVAL} ${PVAL}

# send alert if necessary
FLAG=$(grep -c -e '+' -e '!' ${PVAL})
if [[ ${FLAG} -gt 0 ]]
then
   mailx -s "Filesystem Threshold Exceeded" $RECIPIENTS <<EOF
Date = $(date)
Host = $(hostname)
Threshold = ${PCT}%
Key: = no change; + increase; ! new entry

FileSystem  In Use         Change
==========  ======         ======
$(cat ${PVAL})
EOF
fi

rakesh kumar mall · 09-03-2011, 09:37 AM

You must need to configure nagios. It's good for monitoring CPU , Memory, HDD etc.You can also set e-mail and msg alert in nagios.

digital-infinity · 09-05-2011, 02:04 AM

As root write to the file /etc/cron.d/check_disk_space this:

Code:

0 8 * * 1-5 root /bin/df -k | sed -e 's/\%//g' | /bin/awk '{if (xx>0 && $5>90) print $1" "$5"\%"; xx++;}

This one line

script will check daily at 8:00 whether any mounted file system usage exceeds 90%. It sends email if there is any such file system.