* How can I add start/stop commands for the script.

One way to do it would be to add the respective options to your script (as you're already using Getopt::Long, this should be rather straightforward).  The implementation would typically involve a pid-file to store the PID which you need to communicate with a potentially running instance of the script. Something like:

• start
  

  • Check if there's a pid-file.
  • If so, read the PID from the file, and send that PID signal number zero to check if it's still alive.
  • If it is alive, output "... already running", and quit.
  • In case no running instance is found, start the daemon the same way you would do from the command line. When successful, create a pid-file containing the PID of the process.

• stop
  

  • Check if there's a pid-file.
  • If there is no pid-file, output "... doesn't seem to be running", and quit.
  • Otherwise, read the PID from the file, and send that PID the TERM signal (or whatever signal makes your script quit gracefully).
  • Delete the pid-file. (In case you're paranoid, you might want to wait a while and then check whether the TERM signal did actually terminate the process, before deleting the pid-file. If not, output an error message.)

To avoid code duplication, use a subroutine.

Also you will find your code easier to read an debug if you use consistent indenting. The block of code controlled by a flow-of-control statement should always be indented inside the flow of control statement. It is also a good idea to be consistent in your choice of cuddled and uncuddled curly braces.

A third bit of advice, $@ can sometimes be set to undef even if there is an error. Even so, eval will always return undefined when there is an error. A much better way to check for errors is the incantation:

eval {
  # my stuff ...

  #Note: return *only* exits eval {...}
  #make sure success "returns" true
  return 1;
} or do {
  # what you are currently putting inside if ($@) {...} goes here.
}
[download]

With a subroutine and the above eval incantation, your code (indenting revised):

becomes

sub runMailHandler {
  my ($opt_mailid) = @_;

  my ($log) = Log::Log4perl::get_logger('main');
  my $data = get_data('mail_id' => $opt_mailid
                      , 'file_path' => "/spool/emails");

  ##connect db
  MysqlUtils::connect();

  eval {
    MysqlUtils::insert_record($data);
    return 1;
  } or do {
    $log->error("unable to insert data");
  }    
}

if($opt_daemon) {
  # intro stuff ...
  while (1) {
    runMailHandler($_) foreach @mail_ids;
  }
} else {
  runMailHandler($opt_mailid);
}
[download]

Best, beth

Update Added comments about eval handling.

I prefer daemontools for background processes, as explained in Re: keep a module in shared memory. No need to add a single line of code to your script, no race conditions when handling PID files (because daemontools do not need PID files), and a simple control mechanism. Just write a char to a pipe or let the svc utility do that for you.

(And yes, djb's behaviour regarding errno is just silly and childish. Apply the patch from thedjbway.org or complain loudly to djb that he does not want to accept the C standards. And by the way, also apply the other patch that adds support for SIGQUIT, SIGUSR1, and SIGUSR2.)

Alexander