On Thursday, 12/04/2008 at 06:54 EST, Howard Rifkind <[EMAIL PROTECTED]> wrote: > I'm running exec X on cms user 'A' > > I exec X has to start exec Y running on cms user 'B' and after exec Y > terminates it has to return control to the exec on cms user 'A' > > I could use some suggestions as to the best way to do this.
If you're on the same VM system, REXEC is the best way to do that, assuming that user 'B' is not already logged on. REXEC has all the function to autolog a user, run a command, watch the console for output, send the output back to user A (as output from the 'rexec' command, trappable by a Pipe), and then logoff the user B. Use LOGON BY support so that User A doesn't have to know User B's password, just his own. (You can't do it using rexec without authentication. Well, you can, but then anyone could run the program in User B, not just User A.) If you're not on the same system, life gets more difficult since rexec isn't appropriate for open network connections as it isn't encrypted. One might be tempted to rexec into a local Linux guest that uses ssh to reach the remote system and use rexec over there, piping the response all the way back to you. This all assumes you're not wanting to write a remote execution solution yourself. If you do, just remember that you will need some sort of authentication and authorization. Alan Altmark z/VM Development IBM Endicott
