in reply to Using SIGHUP to restart a daemon

Hack small proof-of-concept code which reproduce your problem, and post here, so other can test it.

Replies are listed 'Best First'.
Re^2: Using SIGHUP to restart a daemon
by ibm1620 (Hermit) on Oct 17, 2013 at 16:25 UTC
    If you follow the link in the original post, it takes you to perlipc doc containing code that illustrates the problem.
      Looks like it's known problem in perl.
      Also, similar issue Can't catch signals after an exec?

      Code can be simplified to
      $sub = sub { exec "perl $0" }; $SIG{INT} = $sub; $SIG{USR1} = $sub; while(1) { print ++$counter, "\n"; sleep 1; }
      (if you press Ctrl-C, it restarts only once)
      Workaround for example in perldoc:
      #!/usr/bin/env perl use POSIX (); use FindBin (); use File::Basename (); use File::Spec::Functions; $| = 1; # make the daemon cross-platform, so exec always calls the script # itself with the right path, no matter how the script was invoked +. print STDERR "STARTED\n"; my $script = File::Basename::basename($0); my $SELF = catfile($FindBin::Bin, $script); # POSIX unmasks the sigprocmask properly my $need_restart = 0; $SIG{HUP} = sub { print "got SIGHUP\n"; $need_restart = 1; }; code(); sub code { print "PID: $$\n"; print "ARGV: @ARGV\n"; my $count = 0; while (++$count) { sleep 2; if ($need_restart) { exec($SELF, @ARGV) || die "$0: couldn't restart: $!"; } print "$count\n"; } }

      (i.e. need to call exec() outside of signal handler)

      And another workaround is posted above by McA Re: Using SIGHUP to restart a daemon