Hello there,

I've ran into the following problem:

I currently have a cron setup that runs every minute. It's basically something like:
Problem is, if NFS fails on otherserver, I get an endless ammount of crond and sh populating the process list which eventually causes the load to increase as high as 80-100 which then causes other problems.

I've tried to fix the problem my mounting like this:
I believed the 20 second timeout would solve the problem but it did not.

Anyone could suggest something else to avoid this problem?

Thank You

You should modify script.sh so that it detects whether a copy is already running and exits if so. Something like this:

#!/bin/bash

if ps -ef |grep script.sh |grep -v grep
then echo ERROR: script.sh already running.
exit 2
fi

...rest of the script...

That way it will abort regardless of the reason it is still running including NFS failure.

IMHO, even though cron can go down to 1 minute intervals, anything more than every 5 mins & I seriously consider writing a daemon instead.
all you need is

nohup infinite_loop_around_code &

Creating a new process every minute is adding to the system load unnecessarily. This also automatically avoids more than one copy running.
I'd also change the nfs params to use soft instead of hard.
You can easily get the daemon to check the mnt before trying to use it.