excellent suggestions by the perlmonks.
an alternative to copying the file in the first place could be to have a perl cgi produce the instruction set when requested, thus presenting it via a web server. then machine 2, or any other machine or machines, simply perform a http get and parse the information into the required schedule and execute. this gets around the whole file permissions and copying mechanism. you could then even extend the perl cgi to be parameter driven, i.e. provide different information depending on parameters passed. if a www server is available or can be easily installed on a machine, it simplifies the whole thing.