but it seems like there is no way to access any of the variables of the parent.

From the on_finish handler? That makes no sense. What did you try?

Or from the child? The only communication P::FM gives you is the exit code. If that's not good enough, you'll need to create your own. (Assuming that's possible.)

Okay, I spoke too soon. Accessing a hash defined in the parent IS read/writable from run_on_finish().

my %rc_hash         = ();
my $MAX_PROCESSES    = 120;
my $pm = new Parallel::ForkManager($MAX_PROCESSES);
    
$pm->run_on_finish(
     sub { 
      my ($pid, $exit_code, $ident) = @_;
         print "run_on_finish: $ident (pid: $pid) exited " . "with cod
+e: [$exit_code]\n";
    $rc_hash{$pid} = $exit_code;
         }
);

$pm->run_on_start(
      sub {
          my ($pid,$ident)=@_;
          print "** $ident started, pid: $pid\n";
       }
 );

for(my $i=0; $i< @hosts; $i++){
          $pm->start($hosts[$i]) and next;   
           system(@some_command);
           $return_code = $? >> 8;
           $pm->finish($return_code);
}

foreach my $key (keys %rc_hash){
      print "$key => $rc_hash{$key}\n";
}
[download]

I made an incorrect assumption.. What I was trying to do earlier is use a database handle from run_on_finish() and that didn't work, so I thought it applied to all other variables. I am not sure why accessing a hash would work, but accessing a db handle wouldn't. Is this possible?

my $dbh    = DBI->connect("DBI:mysql:database=mydb;host=$DB", "user", 
+"pass", {'RaiseError' => 1});
my $MAX_PROCESSES  = 120;
my $pm = new Parallel::ForkManager($MAX_PROCESSES);
    
$pm->run_on_finish(
     sub {
            my ($pid, $exit_code, $ident) = @_;       
            print "run_on_finish: $ident (pid: $pid) exited " . "with 
+code: [$exit_code]\n";
    insert_into_db(\$dbh, $pid, $exit_code);      
+    
          }
);
     $pm->run_on_start(
               sub {
                   my ($pid,$ident)=@_;
                   print "** $ident started, pid: $pid\n";
               }
               );

    for(my $i=0; $i< @hosts; $i++){
          $pm->start($hosts[$i]) and next;   
           system(@some_command);
           $return_code = $? >> 8;
           $pm->finish($return_code);
        }
$dbh->disconnect();

sub insert_into_db{
  my $dbhdl     = shift;
  my $pid       = shift;
  my $ret_code  = shift;

  $$dbhdl->do(INSERT INTO system_results ... .. );
}
[download]

If I can't use the db handle from run_on_finish(), I can still add the return code to %rc_hash and process them after $pm->finish(). Is this the right thing to do? thanks for the speedy responses. I hope I explained my problem a little better this time around.

What I was trying to do earlier is use a database handle from run_on_finish()

The child inherited the database handle and closed it. P::FM uses exit when it should probably use POSIX's _exit. finish is simply a call to exec, so if you call _exec instead of finish, that problem should go away.

That means you should replace

for(my $i=0; $i< @hosts; $i++){
    $pm->start($hosts[$i]) and next;   
    system(@some_command);
    $return_code = $? >> 8;
    $pm->finish($return_code);
}
[download]

with

use POSIX qw( _exit );

for(my $i=0; $i< @hosts; $i++){
    $pm->start($hosts[$i]) and next;   
    system(@some_command);
    $return_code = $? >> 8;
    _exit($return_code);
}
[download]

Of course, system + _exit is a silly way of doing exec, so you get:

for my $host (@hosts) {
    $pm->start($host) and next;   
    exec(@some_command);
}
[download]

Error checking added:

for my $host (@hosts) {
    $pm->start($host) and next;   
    exec(@some_command);
    print(STDERR "exec failed: $!\n");
    _exit($!);
}
[download]

That also fixes the bug where you'd return zero (success) when the child died from a signal.

Update: Clean up.

Use a closure:

my $stats = [];
my $code  = create_closure( $stats );

....

$pm->run_on_finish( $code );

....

print Dumper( $stats );


sub create_closure {
  my( $stats ) = @_;
  return sub {
    my ($pid, $exit_code, $ident) = @_;
    push( @$stats, { 
      pid       => $pid, 
      exit_code => $exit_code, 
      ident => $ident 
    } );
  }
}
[download]

He already is, which is why I asked to see the problem.

Unless the OP edited the code, his example does not have a closure. He alluded to tracking the exit codes via a HASH but that sounded more like a global than a closure. I believe the OP may be looking for someway to update his database inside the run_on_finish command -- in that case he could just close over the database handle instead of some data structure which could be manipulated *after* the run of all the children.

Update: ... forgot an (untested) example:

my $dbh = create_dbh_handle();
my $code = create_closure( $dbh ); 
.... 
my $pm = Parallel::ForkManager->new( $MAX_NR_PROCESSES );
   $pm->run_on_finish( $code ); 

foreach ... {
  my $pid = $pm->start and next;
  ...
  $pm->finish; # Terminates the child process
}

sub create_closure { 
  my( $dbh ) = @_; 
  return sub { 
    my ($pid, $exit_code, $ident) = @_; 
    my $sql = "insert into table values( $pid, $exit_code )";
    $dbh->do( $sql );
  }
}
[download]

-derby