It all depends on what you are doing. It really sounds like you need to write a shell script and use scriptutil to execute it. Are your end targets Unix based servers?
Why do the for on the box, w/ nsh you can do things like:
for server in $serverlist
nexec $server sh -c "<some command>"
foo=`nexec $server sh -c "<some other command>"
if [ $foo = bar ]
If you want to execute a simple command, then just enter the command you want run under "Command / Script" and select "Remote Execution" as the "Execution Type". The command will be spawned out on the target system. This method does not support pipes (e.g. cmd1 | sort) and, I suspect, would probably not deal well with shell metacharacters (wildcards, parentheses, etc).
If you need multiple commands, then put it in a shell script file. You can use scriptutil, runscript, or nsh -c in the "Command / Script" field to make your script run out on the target system. You will not be mixing 'nexec' into your script; your entire script is exec'ed out onto the system.
It seems to me almost as if you're taking the initial concept of being able to use nexec to run small bits out on remote systems and trying see Extended Objects as centralized code that can do the same thing. Instead, think of Extended Objects as a library of code that can be copied out to local contexts and run locally, with results returned to the center.
Does that help any?
thanks guys, all your comments where very helpful and solved the issue I had.