Here is a simple shell script tested on CentOS / RHEL / Fedora / Debian / Ubuntu Linux. Should work under any other UNIX liker operating system. It will check for httpd pid using pgrep command
pgrep command
pgrep looks through the currently running processes and lists the process IDs which matches the selection criteria to screen. If no process found it will simply return exit status 0 (zero).
Download the script and set cronjob as follows:
*/5 * * * * /path/to/script.sh >/dev/null 2>&1
Sample script
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 | #!/bin/bash # Apache Process Monitor # Restart Apache Web Server When It Goes Down # ------------------------------------------------------------------------- # Copyright (c) 2003 nixCraft project <http://cyberciti.biz/fb/> # This script is licensed under GNU GPL version 2.0 or above # ------------------------------------------------------------------------- # This script is part of nixCraft shell script collection (NSSC) # Visit http://bash.cyberciti.biz/ for more information. # ------------------------------------------------------------------------- # RHEL / CentOS / Fedora Linux restart command RESTART="/sbin/service httpd restart" # uncomment if you are using Debian / Ubuntu Linux #RESTART="/etc/init.d/apache2 restart" #path to pgrep command PGREP="/usr/bin/pgrep" # Httpd daemon name, # Under RHEL/CentOS/Fedora it is httpd # Under Debian 4.x it is apache2 HTTPD="httpd" # find httpd pid $PGREP ${HTTPD} if [ $? -ne 0 ] # if apache not running then # restart apache $RESTART fi |
A better and more reliable solution is monit monitoring software for restarting services such as mysql, apache and sendmail under UNIX / Linux operating systems.
27 comment
Why won’t you just use ‘monit’?
I am not a big fan of greping httpd in process list to see if httpd server is running or not. There can be several httpd services installed. Two options are either the status module, or this bash script to the extent that you known the full path to your httpd bin folder:
function isapup(){
pid=$($1/bin/apachectl -S | grep PidFile | awk ‘{print $2}’) # retrieve the expected path of the process id file
pid=${pid:1:
expr ${#pid} - 2} # get rid of the quotesup=$(ls $pid 2> /dev/null) # see if process id file exists
if [ ! “$up” ]
then
return 1
fi
}
isapup /opt/httpd/2.4.0
if [ $? -eq 1 ] # if isapup returned 1
then
echo Starting Apache
/opt/httpd/2.4.0/bin/apachectl start
fi
Had to use PGREP=”/usr/bin/pgrep -l -x” to get it to work on CentOs (5.4)
If adapting this for other processes add an -x flag to force pgrep to match the process name exactly
$PGREP ${HTTPD}
The name of my script contains the name of the the process to pgrep was giving a false positive when the process wasn’t running.
You made mistake in crontab entry. It will not work.
You should specify user which will execute the script:
This crontab is correct (CentOS):
*/5 * * * * root /path/to/script.sh >/dev/null 2>&1