Shell Programming and Scripting

Hi all,

My tiny mind is struggling with the following script that shuts down Tomcat on Sol. 8 and kills any defunct processes Tomcat might leave behind (as it so often does). I realize that fixing Tomcat would be best, but this is a band-aid so I won't have to do as much off-hours support of this pig, which will hopefully buy me time to research the matter more thoroughly.

Anyway, on to the goods...

Code:

#!/bin/sh
# variables
tcpid=`ps -ef |grep java |grep -v grep |sort |awk '{print $2}'`
signal=9
tclog=/opt/local/jakarta/logs/catalina.out
now=`date +'%Y%m%d%H%M'`

# Shutdown Tomcat
/opt/local/jakarta/bin/shutdown.sh

sleep 20

if [ $tcpid != "" ]
do
    (sudo kill $signal) $tcpid
        sleep 5
else [ $tcpid = "" ]
        echo "Java Stopped"
done

# manage catalina.out, cache dirs and core dumps
cp $tclog /opt/local/jakarta/logs/catalina.out.$now
>/opt/local/jakarta/logs/catalina.out
gzip /opt/local/jakarta/logs/catalina.out.$now
rm -rf /opt/local/jakarta/work/Apache/*
find /opt/local/jakarta/ -name core |xargs rm
# wait for tomcat cache to clear
sleep 30

# restart this pig
/opt/local/jakarta/bin/startup.sh

exit

So far, my problems are in this segment:

Code:

if [ $tcpid != "" ]
do
    (sudo kill $signal) $tcpid
        sleep 5
else [ $tcpid = "" ]
        echo "Java Stopped"
done

I want to see if the pid still exists (thus is defunct, needing sudo to kill), then kill -9 it every 5 seconds or so until it goes away. Once it's gone, it can move on to the rest of the script.

What this one's doing is looping infinitely _and_ I keep getting errors about it trying to kill a process called '-9'. I've tried

`sudo kill -9 $tcpid` or (sudo kill -9 $tcpid), but to no avail.

Anyway, any pointers for how I can effectively pull this off?

Thanks!

Dave

I've tried this type of while loop too (since I think a while loop makes more sense for what I'm trying to do):

Code:

while [ $tcpid != "" ]
do
    (sudo kill -9 $tcpid)
    sleep 5
done

It runs regardless of the whether $tcpid is running or not, and it too loops until I manually kill it.

I see in my script above my signal variable was only '9' and not '-9'. I've tried both ways and that doesn't work either.

Let's say that the pid is 1612. So you set tcpid to 1612 and then you enter your loop. On the first iteration you kill the process. After the process dies, the value of $tcpid will not magically change values for you. It is still 1612. It will stay 1612 forever since you don't assign it a new value in the loop.

What is the point of the loop anyway? If a process survived a "kill -9", do you think that second "kill -9" will work any better?

Thanks for the reply. I'll try it with the full command (thus re-evaluating whether the pid exists) to see if that helps.

Regarding your question, the process may not sustain a kill -9 over a prolonged period of time, but it takes so long for it to stop, that I've found just spamming signals its way kills its defunct process. The goal is to bring the process down and up as soon as possible.

Cheers,

Dave