We (my work) drive our functional tests for a distributed system using
Ruby scripts.  It seems like every line is a back-tick to ssh to do
something interesting on some other box.  Works like a charm.  Just
have to remember to check for errors, as neither system nor backtick
automatically do anything interesting when the shell command fails.
We usually end up using a little function that does whatever backticky
thing we want and throws an exception if it doesn't work -- then we
can stop having to check $? all the time.

We use key exchanges, as Patrick suggested, as well as sudo privileges
on the remote boxes (or just ssh in as root.  Sometimes laziness wins
in the battle to get a functional test written).

  Wayne Conrad

On Fri, May 17, 2002 at 02:42:51PM +0900, Patrick May wrote:
> Todd Holloway <todd / duckland.org> wrote in message news:<20020516190738.GA27793 / duckland.org>...
> > I'd like my first Ruby project to be simple user-mangement program
> > which uses ssh (preshared key part is already working).
> > 
> > Has anyone written anything close or does anyone have any suggestions?
> > I'd eventually like to have it grant "sudo" permissions as well...
> > and longterm store all in an SQL database.
> 
> ruby supports the `shell command here` syntax.  This could be a quick
> way of connecting the ruby script to the commands you want it to
> execute.