I have a very simple backup script. When I run it from command line it works well, so I created a symbolic link to it in cron.daily.

The script is executed, but a part of it, that calls the backup application for the Firebird database doesn't work.

I think it is something with permissions, since the backup runs fine when I'm looged as root.

Scripts executed by cron runs under a specific account? How does it work? Where can I find more information about it?

You may have to set the paths to search for executables in the script since this will most likely not be set, or you can specify the full path to all programs called from the script (instead of mybackup.sh use /usr/local/bin/mybackup.sh).


Håkan

That's it! Thanks!

Since I can start the application from anywhere while I'm in console, I thought I didn't need to specify the full path in the script.

But suppose I want the scripts to use the search paths, how can I do it?

You need to set the path at the beginning of your script. The environment variable $PATH contains a comma-separated list of the directories that are to be searched for executables. Type echo $PATH to see what directories are currently in your $PATH.

You set it like this:
PATH=dir1:dir2:dir2
...or for a more real-world like example:
PATH=/bin:/usr/bin:/usr/local/bin:~/bin

You can also tack on a directory to the existing $PATH by including $PATH itself in the list of directories to be used:
PATH=$PATH:/opt/myownstuff/bin:/home/jenny/bin

Simply adding the PATH= entry with the directories you need at the top of the script, after the magic line (#!/bin/bash or whatever) but before any real stuff gets done should let you use the programs in the directories you specified without using their full paths.

On a related but not really relevant note, I usually use this in my ~/.bashrc:
alias path='echo -e ${PATH//:/\\n}'
This provides me with the command path that will print the current $PATH in a more readable format - one directory per line.

Håkan

Thanks a lot!