signals in 2.0

[Ryan Bloom <rb...@ntrnet.net>]

I have been playing with the prefork MPM recently, and it turns out that
prefork uses SIGUSR1 for graceful shutdown.  Any threaded MPM is using
SIGWINCH for graceful shutdown because Linux threads monopolize SIGUSR1
and SIGUSR2 for any 2.0 kernel.  What I am wondering, is if we really want
different signals depending on which MPM is used.

Ryan


Come to the first official Apache Software Foundation
Conference!!!   <http://ApacheCon.Com/>

_______________________________________________________________________________
Ryan Bloom                        	rbb@ntrnet.net
2121 Stonehenge Dr. Apt #3
Raleigh, NC 27615		Ryan Bloom -- thinker, adventurer, artist,
				     writer, but mostly, friend.
-------------------------------------------------------------------------------

Re: signals in 2.0

[Manoj Kasichainula <ma...@io.com>]

On Thu, Feb 10, 2000 at 02:02:31PM -0500, Ryan Bloom wrote:
> What I am wondering, is if we really want different signals
> depending on which MPM is used.

No, we don't, or at least we don't want the defined interface for
shutdown/restart to be different between MPMs (or even platforms). The
problem is that Linux+threads can't use SIGUSR1, and I believe that
there are old platforms that can't use SIGWINCH.

There are a few possibilities I've been able to think of so far (from
easiest to hardest):

- screw Linux 2.0, and say that we use SIGUSR1 for graceful restart
  everywhere. All this would mean is that Linux 2.0 users would have
  to use a preforking MPM.

- Add -k options under Unix, like the ones in Windows, that will send
  whatever the defined signal is for that platform+MPM.

- Eliminate non-graceful restart, and use SIGHUP for graceful restart.
  People who want the non-graceful restart can SIGTERM and start the
  server again.