watchdog.sh script for checking server running not mine, i stole it)

watchdog.sh

#!/bin/bash
# Service watchdog script
# Put in crontab to automatially restart services (and optionally email you) if they die for some reason.
# Note: You need to run this as root otherwise you won't be able to restart services.
#
# Example crontab usage:
#
# Strict check for apache2 service every 5 minutes, pipe results to /dev/null
# */5 * * * * sh /root/watchdog.sh apache2 "" > /dev/null
#
# "Loose" check for mysqld every 5 minutes, second parameter is the name of the service
# to restart, in case the application and service names differ. Also emails a report to [email protected]
# about the restart.
# */5 * * * * sh /root/watchdog.sh mysqld mysql [email protected] > /dev/null
 
# Common daemon names:
# Apache:
# apache2 - Debian/Ubuntu
# httpd - RHEL/CentOS/Fedora
# ---
# MySQL:
# mysql - Debian/Ubuntu
# mysqld - RHEL/CentOS/Fedora
# ---
# Service name
DATE=`date +%Y-%m-%d--%H-%M-%S`
SERVICE_NAME="$1"
SERVICE_RESTARTNAME="$2"
EXTRA_PGREP_PARAMS="-x" #Extra parameters to pgrep, for example -x is good to do exact matching
MAIL_TO="$3" #Email to send restart notifications to
  
#path to pgrep command, for example /usr/bin/pgrep
PGREP="pgrep"
 
#Check if we have have a second param
if [ -z $SERVICE_RESTARTNAME ]
  then
    RESTART="/sbin/service ${SERVICE_NAME} restart" #No second param
  else
    RESTART="/sbin/service ${SERVICE_RESTARTNAME} restart" #Second param
  fi
 
pids=`$PGREP ${EXTRA_PGREP_PARAMS} ${SERVICE_NAME}`
 
#if we get no pids, service is not running
if [ "$pids" == "" ]
then
 $RESTART
 if [ -z $MAIL_TO ]
   then
     echo "$DATE : ${SERVICE_NAME} restarted - no email report configured."
   else
     echo "$DATE : Performing restart of ${SERVICE_NAME}" | mail -s "Service failure: ${SERVICE_NAME}" ${MAIL_TO}
 fi
else
  echo "$DATE : Service ${SERVICE_NAME} is still working!"
fi

# copylefted from https://gist.github.com/vodolaz095/5073080

watchdogMaster.sh

#!/bin/bash
mailto="mymail@mydomain"

/bin/bash /root/watchdog.sh mysqld mysqld "$mailto"
/bin/bash /root/watchdog.sh httpd httpd "$mailto"
/bin/bash /root/watchdog.sh pound pound "$mailto"
/bin/bash /root/watchdog.sh redis-server redis "$mailto"
/bin/bash /root/watchdog.sh memcached memcached "$mailto"
/bin/bash /root/watchdog.sh searchd searchd "$mailto"

@zednight, i'm surprised for so many comments for script of unknown authorship or origin i manage to dig in HDD from server that stopped working in 2005 year....
Are you sure that ps -e shows exactly the command for watchdog.sh?
my PC have few watchdogs already, not related to this script

[vodolaz095@steel ~]$ ps -e | grep watchdog
   12 ?        00:00:00 watchdog/0
   13 ?        00:00:00 watchdog/1
   20 ?        00:00:00 watchdog/2
   27 ?        00:00:00 watchdog/3

For Ubuntu16.04 - Linux

RESTART="/sbin/service
change to
RESTART="/usr/sbin/service

and

*/5 * * * * bash /root/watchdog.sh mysqld mysql [email protected] > /dev/null
*/5 * * * * bash /root/watchdog.sh apache2 "" [email protected] > /dev/null

Thanks!
just one more question, if I send STOP siginal to the process id, then the STAT of the process changed from S=>T, how to check?

Thanks for sharing @vodolaz!
I use this script to restart the memcached service if for some reason this service has failed.

Mine is a bit shorter...

#!/bin/sh
app='[some command to start your stuff]'
if [ ! "$(pidof $app)" ]
then
[some command to start your stuff]
fi