Shell Programming and Scripting

Hello, I'd like to create a small application which would monitor one important log on our Solaris 10 environment.
So far, I createad a script which does the monitoring, and I though through nohup it would become one of the permanent running processes.

>nohup monitor.sh &
[1] 3139
> more monitor.sh
#!/bin/ksh
LOGDIR=/application/logs
if [ ! -f ${LOGDIR}/extracted_errors.log ]
then touch ${LOGDIR}/extracted_errors.log
fi

tail -f application_common.log |
while read -r line
do
echo ${line} | grep 'ALERT' | grep 'first_application' | grep 'SourceError' >> /dev/null
if [ $? = 0 ]
then
echo "Logged at:"`date '+%h %d %H:%M:%S'`"1st Application failure" >> ${LOGDIR}/extracted_errors.log

echo ${line} | grep 'ALERT' | grep 'second_application' | grep 'SourceError' >> /dev/null
if [ $? = 0 ]
then
echo "Logged at:"`date '+%h %d %H:%M:%S'`"2nd Application failure" >> ${LOGDIR}/extracted_errors.log
fi
done

Now, my problem is, it runs and does what I want, but how can I stop it?
When I want to exit the system (and leave the monitoring script running), it's not allowing me to leave. The only way how to leave, is to kill the tail -f process for other terminal connection
What am I doing wrong?
Any help appreciated...
Thank you in advance.

Martin

The problem is the "tail -f". This ties you down.

Try using "wc -l" on the log and keeping a record in a "previous lines count" file. Then sleep and calculate the number of lines by which the log has grown and "tail -<lgrowth>" the log if it has grown. Then save the current line count, sleep for a period and repeat the test. Within this loop test for a "time to exit" file and exit when this file exists.
Obviously such a script needs a "first time" condition.

Hope this makes sense.

Hmm cfajohnson. How does the o/p stop the process without using kill ?

Thank you guys for your help.
wc -l wouldn't work too well, because old datas from the log are backed up (so there would have to be some IFs...)
Meanwhile I had a lot of other job and now when I came back to this problem I found a solution, I created a small stop.sh, which adds some texts like 'You should STOP' at the end of application_common.log, and then let the main script in while statement do following test:

echo ${line} | grep 'You should STOP' >> /dev/null
if [ $? = 0 ]
then
echo "Logged at:"`date '+%h %d %H:%M:%S'`"The monitor is being shut down" >> ${LOGDIR}/extracted_errors.log
exit 0
fi

This works fine.. so far
So if anybody would have the same problem, this could be a soultion for you too.