Very useful in shell scripts because you can run a task nicely in the background using job-control and output progress until it completes. Here's an example of how I use it in backup scripts to run gpg in the background to encrypt an archive file (which I create in this same way). $! is the process ID of the last run command, which is saved here as the variable PI, then sleeper is called with the process id of the gpg task (PI), and sleeper is also specified to output : instead of the default . every 3 seconds instead of the default 1. So a shorter version would be sleeper $!; The wait is also used here, though it may not be needed on your system. echo ">>> ENCRYPTING SQL BACKUP" gpg --output archive.tgz.asc --encrypt archive.tgz 1>/dev/null & PI=$!; sleeper $PI ":" 3; wait $PI && rm archive.tgz &>/dev/null Previously to get around the $! not always being available, I would instead check for the existance of the process ID by checking if the directory /proc/$PID existed, but not everyone uses proc anymore. That version is currently the one at http://www.askapache.com/linux-unix/bash_profile-functions-advanced-shell.html but I plan on upgrading to this new version soon. Show Sample Output

Referring to the original post, if you are using $! then that means the process is a child of the current shell, so you can just use `wait $!`. If you are trying to wait for a process created outside of the current shell, then the loop on `kill -0 $PID` is good; although, you can't get the exit status of the process.

Ssh to host1, host2, and host3, executing on each host and saving the output in {host}.log. I don't have the 'parallel' command installed, otherwise it sounds interesting and less cryptic.

sleep in microseconds instead of seconds Alternatively to usleep, which is not defined in POSIX 2008 (though it was defined up to POSIX 2004, and it is evidently available on Linux and other platforms with a history of POSIX compliance), the POSIX 2008 standard defines nanosleep Show Sample Output

If you really _must_ use a loop, this is better than parsing the output of 'ps': PID=$! ;while kill -0 $PID &>/dev/null; do sleep 1; done kill -0 $PID returns 0 if the process still exists; otherwise 1

This command explains how to manage some asynchronous PID in a global process. The command uses 4 processes in a global process. The asynchronous scripts are simulated by a time.sh script more infos : http://code-esperluette.blogspot.fr/2012/03/bash-gestion-de-processus-asynchrones.html http://www.youtube.com/watch?v=TxsPyAtD70I

Show Sample Output

The '[r]' is to avoid grep from grepping itself. (interchange 'r' by the appropriate letter) Here is an example that I use a lot (as root or halt will not work): while (ps -ef | grep [w]get); do sleep 10; done; sleep 60; halt I add the 'sleep 60' command just in case something went wrong; so that I have time to cancel. Very useful if you are going to bed while downloading something and do not want your computer running all night.