Rerun a cronjob if it doesn't work at first.

I want to add a line to the script that searches for a file. If the file does not exist, I need to kill the cronjob and run it again in one hour. I am running the program and 2 am and 2pm. I would want the job to run again at 3am and 3pm respectively. How could I do this and still keep the cronjob running twice daily if I killed it once? The cronjob simply runs one file. example

sh filebatch.sh

. Then filebatch.sh runs several other files, example

 $sh filebatch2.sh $sh filebatch3.sh $sh filebatch4.sh

. Filebatch2.sh is a script that uses wget to pull files that I need to run the program. If the files are not ready, I want it to kill the entire job and run it again in 1 hour. I would like to put the line of code right after the command

 wget file1 

. How would I do this. I think I am using a linux machine with a c shell. See below.

[rhuff@huina working]$ echo $SHELL
/bin/csh

---------- Post updated at 01:18 PM ---------- Previous update was at 12:09 PM ----------

Looks like I also have bash installed on the box. It looks like when I use an sh command it runs bash

Cron is a very basic scheduler without any real conveniences such as retries, so you'll need to get creative.

My first thought is to schedule the script at both times (2pm and 3pm for instance) and have the first script delete, or move, the input file. The second execution will fail because the first execution processed the file, and if the file wasn't there for the first execution, the second should find it. If the script is already deleting/moving the file, then no changes will be needed.

If you need to keep the file, another simple solution would be to have the script sleep for an hour and retry if it doesn't find the file initially. Depending on the complexity of your script this will require the least amount of change to your script.

The only other way that I can think of is a bit of a kludge, and will require a few modifications to your script, but it will work.

First, modify your script to create a 'rerun' file in /tmp or some other known location when it is not able to find the file that it needs. It then exits.

A second modification, to accept a check option from the command line is also necessary. If the check option is present, then the script should initially test for the 'rerun' file. If the rerun file isn't present, then the script should immediately exit. The fact that there is no rerun file indicates that the earlier execution was successful.

Once those changes are made, add two more lines to your crontab that invoke the script at 3pm and 3am with the check option.

A small (untested) example of this (ksh/bash Im not a Csh person at all):

#!/usr/bin/env ksh

create_rerun=1            # default to creating rerun on failure
if [[ $1 == -c ]]
then
   if [[ ! -f /tmp/my.rerun ]]
   then
       echo "did not find rerun file; exiting OK"
       exit 0
    fi

   create_rerun=0     # don't create a rerun on failure 
fi

rm -f /tmp/my.rerun            # ensure that the rerun file is gone
if [[ ! -f needed_file_name ]]
then
   echo "unable to find needed file"
   if (( $create_rerun > 0 ))
   then
      touch /tmp/my.rerun
      exit 0                       # stop now, an ok return code
   fi
   exit 1                 # stop now -- error
fi

# what ever code you have to manipulate the file

If the script cannot be modified, then create a small 'wrapper' that is basically the code above and invoke the wrapper from cron. The wrapper then invokes the real script, but only if the file exists.

The crontab entries would be something like this for the check method:

0 2 * * * /usr/home/me/bin/myscript 
0 3 * * * /usr/home/me/bin/myscript -c