Automated Script for Process killing

Hello all,

I'm in need of a Script which needs to wait for all the child process to end and then kill the main process.

I have a process called

mainpp

which runs for different instances like

evpn, nge, gmn etc 

so when i query for

mainpp

process it looks like below.

bash-3.2$ ps -eaf | grep -i mainpp
   mainpp 14474 13672   0   mrt 08 ?          14:17 /opt/mainpp/bin/mon /opt/mainpp/etc/emtn-gsm.conf
   mainpp 14535 13672   0   mrt 08 ?          17:25 /opt/mainpp/bin/mon /opt/mainpp/etc/gmnjuniper.conf
   mainpp 14627 13672   0   mrt 08 ?           6:28 /opt/mainpp/bin/mon /opt/mainpp/etc/pwifi.conf
   mainpp 14636 13672   0   mrt 08 ?        1109:43 /opt/mainpp/bin/mon /opt/mainpp/etc/wwp.conf
   mainpp 14546 13672   0   mrt 08 ?          31:07 /opt/mainpp/bin/mon /opt/mainpp/etc/gmnwwp.conf
   mainpp 14440 13672   0   mrt 08 ?        1741:35 /opt/mainpp/bin/mon /opt/mainpp/etc/alcatel.conf
   mainpp 14447 13672   0   mrt 08 ?          13:18 /opt/mainpp/bin/mon /opt/mainpp/etc/asr9001.conf

the child process gets invoked for every 5 mins, and looks like below

bash-3.2$ ps -eaf | grep -i mainpp | grep -i rtr
   mainpp 17628 13672   0 11:10:08 ?           0:00 /opt/mainpp/bin/rtr poll.5 /opt/mainpp/etc/alcatel.conf
   mainpp 17213 13672   0 11:10:01 ?           0:00 /opt/mainpp/bin/rtr poll.3 /opt/mainpp/etc/os6250.conf
   mainpp 17487 13672   0 11:10:06 ?           0:00 /opt/mainpp/bin/rtr poll.10 /opt/mainpp/etc/eok.conf
   mainpp 17211 13672   0 11:10:01 ?           0:00 /opt/mainpp/bin/rtr poll.4 /opt/mainpp/etc/os6250.conf
   mainpp 17025 13672   0 11:10:01 ?           0:00 /opt/mainpp/bin/rtr poll.1 /opt/mainpp/etc/axxerra.conf
   mainpp 17503 13672   0 11:10:06 ?           0:00 /opt/mainpp/bin/rtr poll.0 /opt/mainpp/etc/wwp.conf
   mainpp 17217 13672   0 11:10:01 ?           0:00 /opt/mainpp/bin/rtr poll.1 /opt/mainpp/etc/os6250.conf
   mainpp 17470 13672   0 11:10:05 ?           0:00 /opt/mainpp/bin/rtr poll.4 /opt/mainpp/etc/wwp.conf
   mainpp 17493 13672   0 11:10:06 ?           0:00 /opt/mainpp/bin/rtr poll.7 /opt/mainpp/etc/eok.conf
   mainpp 17483 13672   0 11:10:06 ?           0:00 /opt/mainpp/bin/rtr poll.12 /opt/mainpp/etc/eok.conf
   mainpp 17626 13672   0 11:10:08 ?           0:00 /opt/mainpp/bin/rtr poll.6 /opt/mainpp/etc/alcatel.conf
   mainpp 17457 13672   0 11:10:05 ?           0:00 /opt/mainpp/bin/rtr poll.18 /opt/mainpp/etc/eok.conf
   mainpp 17467 13672   0 11:10:05 ?           0:00 /opt/mainpp/bin/rtr poll.13 /opt/mainpp/etc/eok.conf
   mainpp 17141 13672   0 11:10:01 ?           0:00 /opt/mainpp/bin/rtr poll.0 /opt/mainpp/etc/gmncisco.conf
   mainpp 17465 13672   0 11:10:05 ?           0:00 /opt/mainpp/bin/rtr poll.14 /opt/mainpp/etc/eok.conf
   mainpp 17640 13672   0 11:10:08 ?           0:00 /opt/mainpp/bin/rtr poll.2 /opt/mainpp/etc/alcatel.conf
   mainpp 17633 13672   0 11:10:08 ?           0:00 /opt/mainpp/bin/rtr poll.4 /opt/mainpp/etc/alcatel.conf
   mainpp 17463 13672   0 11:10:05 ?           0:00 /opt/mainpp/bin/rtr poll.15 /opt/mainpp/etc/eok.conf
   mainpp 17501 13672   0 11:10:06 ?           0:00 /opt/mainpp/bin/rtr poll.1 /opt/mainpp/etc/wwp.conf
   mainpp 17449 13672   0 11:10:05 ?           0:00 /opt/mainpp/bin/rtr poll.7 /opt/mainpp/etc/wwp.conf
   mainpp 17511 13672   0 11:10:06 ?           0:00 /opt/mainpp/bin/rtr poll.3 /opt/mainpp/etc/eok.conf
   mainpp 17635 13672   0 11:10:08 ?           0:00 /opt/mainpp/bin/rtr poll.3 /opt/mainpp/etc/alcatel.conf

after 5 mins, the

rtr 

process gets ended on its own depends of work stack it has in its list, once the

rtr

is ended then i would like to kill

mainpp

i need a script which will wait for the

rtr

process to end and then kill the

mainpp

its just one time activity everyday, as i need to reboot the interfacing application and there is no particular time for the same so i need the script to trigger whenever i want.

Please help me with this.

Thanks!!

Do you have pgrep|pkill?

#!/bin/sh
# wait for the rtr processes to terminate
while pgrep -u mainpp -x rtr
do
  sleep 5
done > /dev/null
# terminate the mon processes
pkill -u mainpp -x mon

Otherwise please state your OS

uname

Hello,

Thanks for the solution!!

its not always 5 Sec that rtr takes to complete, its always work stack in its memory that matters.

Apparently, the below script only waits for 5 sec and kill the mainpp abruptly, which i dont want to do

I have written below script to my knowledge

#!/bin/bash
# Version 1.0
# Script to kill all mainpp processes

HOSTNAME=`hostname`
echo $HOSTNAME
BASEDIR=/opt/mainpp
PROCESS_DATA=/opt/mainpp/log/automation_scripts/logs/mainpp_kill_`date +%d%m%Y`
dat=`date +%d%m%Y`
chmod 755 `echo $0`
echo $dat
flag=5

rtr_count=`ps -eaf | grep mainpp | grep -i rtr | grep -v grep | wc -l`

while [ $flag -eq 5 ]
do
  #sleep 30
	if [ $rtr_count -eq 0 ]
	then
		echo "no rtr process is running, mainpp will be killed!! `date`" >> $PROCESS_DATA
		PID=`ps -eaf | grep mainpp | grep -v grep | awk '{print $2}'`
			if [[ "" !=  "$PID" ]]; then
				echo "killing $PID" >> $PROCESS_DATA
			kill $PID
		fi
		break 
	else 
		sleep 30		
	 
    fi
done
exit

Can you please let me know if there are any flaws before i execute the same in prod??

Thanks,

I would think that this is why people have dev environments.
You might want to find a dev environment and try it there first.

Thats true, unfortunately its a standalone system with no dev or acceptance env

Try changing to send an email when it thinks the child processes
are finished. and you can check and see if you agree. Once you
get that working you can change it to find and kill the right
process. You should have the email include everything that you
would need to kill the parent process, then you will know that
the script would have killed the right process.

Your script has an unsharp grep.
For example it will also kill a process /data/domainppd because it contains mainpp .
And what is chmod 755 `echo $0` for??