Repeat job every "m" mins for "h" hrs starting at "t&qu 1

[blodwyn]

Hi

I've always had a very helpful response from you experts out there and I hope that somebody will assist me now.

I'm trying to write a script that will kick off a monitoring job script at a prompted time, to repeat every so many prompted minutes, until a certain time, or for a prompted number of hours.

Because the parameters can vary,and the job is an ad hoc one, I don't want to use the crontab facility. I'm looking at "at", "repeat" and "sleep' but am having trouble loading all the variables into the script. I'm also sure there must be a less messy way.

Thanks in advance

Blodwyn

 

[PHV]

Post the script you have so far and explain where you are in trouble with.

 

[Ygor]

Some ideas that might help...

#!/usr/bin/ksh

read START?'Start time (HH:MM) ? '
read REPEAT?'How often to repeat (minutes) ? '
read END?'End time (HH:MM) OR how long to run (hours) ? '

at $START <<!!
while [ -f flagfile ]
do
run_your_script_here
sleep $((REPEAT * 60))
done
!!

if [[ $END = *:* ]]
then
if [[ $END > $START ]]
then
echo &quot;rm flagfile&quot; | at $END
else
echo &quot;rm flagfile&quot; | at $END tommorrow
fi
else
echo &quot;rm flagfile&quot; | at $START + $END hours
fi

 

[blodwyn]

Thanks Ygor, that looks great - some of the syntax is new to me, which is good.

Where does flagfile come from?

Regards

Blodwyn

 

[Ygor]

I missed that bit out. You will need to put in some validation checks on user input and if all OK then touch flagfile else exit.