The prototype concept is that a list of hosts is stored in a text file. A perl script reads in the hosts list, splits it up among a given number of forked processes, and each process gets to work on pinging. All results are appending into a results file for now.

In essence, I seem to be having trouble coding a fork routine that lives within a loop. Everything grows exponentially. For instance, a file with 10 hosts in it ends up with 212 lines in the result file - the 10 hosts are pinged a total of 212 times. I envisioned that once is enough ;)

My question is, how can I create a variable number of child processes, each chewing on a portion of the data? My code fails at that.

So here's my code. It works great, but not to spec. It is also really bad code, may have blatant errors in implementation, and may be poorly structured. It certainly has not evolved to the point where it waits for child processes to end - a very bad habit, I'm sure. My excuse is that I'm trying (and failing) to prove a concept. My apologies are offered in advance.

#!/usr/bin/perl -w
use strict;

my @hosts = <>; # list of hosts to ping
my $depth = 5; # number of children
my $size = 800; # packet size
my @pid; # store child pids here

open (OUT, ">>results");
for (0..$depth - 1) {
  $pid[$_] = fork;
  if ($pid[$_]) {
    print "$_\n"; # for debugging
  } else {
    foreach my $host (splice(@hosts, $_ * (scalar @hosts)/$depth, ($_ 
++ 1) * (scalar @hosts)/$depth)) {
      chomp $host;
      my @result = `ping -c 1 -s $size $host`;
      $result[1] =~ s/^.*=//mg; # strip the data down to
      $result[1] =~ s/ ms.*$//mg; # decimal miliseconds
      print OUT "$host\t$result[1]";
    }
  }
}
close OUT;
[download]

su-2.04# ./pingfork.pl testping  | sort | uniq -c
  16 0
  16 1
  16 2
  16 3
  16 4
su-2.04# cat results | wc -l
     212