There is on error in your code: sleep uses seconds as an argument. Until you want to wait forever you should choose a smaller value.
The way you handle your processes is a little bit awkward since you do not have any control over them after the forks are run. I cam up with a little script like this:

#!/usr/bin/perl -w

use strict;

sub sched{

  my $file = shift || die "Missing parameter for subroutine call\n";
  my $time = shift || die "Missing second parameter on subroutine call
+\n";

  my $cmd = `cat $file`;

  while (1){
    print $cmd;
    sleep $time;
  }
}

# main program

my $pid; # process id
my @childs; # will hold all pids of our children

$SIG{CHLD} = sub{wait};  # avoid Zombies
$SIG{TERM} =
  sub{
        kill -9, @childs;
    exit(0)
  }; # allow to end program by pressinc Ctrl-C or a kill call

#this will hold all of our tasks
my @tasks = (
  {
    file => "msg.1",
    time => 10,
  },
  {
    file => "msg.2",
    time => 2,
  },
  {
    file => "msg.3",
    time => 5,
  },
);

# looping through the array of hashes and forking with each element

FORLOOP: for my $i ( 0..$#tasks){

         FORK: {
           if ($pid = fork){            # first fork
             # we are parent
             push @childs,$pid;
             next FORLOOP;
           }elsif (defined $pid){       # so we are a child
             sched($tasks[$i]{file},$tasks[$i]{time});
             exit(0); # we should never reach this
           }elsif ($! =~/No more process/){
             sleep 5;
             redo FORK;
           }else{
             die "Something is rotten in the state or Denmark: $!\n";
           }
         }

  }
  print "started all childs\n";
  while (1){} # necessary so we can kill all of our childs
[download]

The hash is just a convenient way to handle all of your files, maybe you want to read those values from a file later. Also note the signal handling which enables you to stop the beasts with a control-c or a kill call.

Hope this helps,
C-Keen