Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender: cygwin-developers-owner AT sourceware DOT cygnus DOT com
Delivered-To: mailing list cygwin-developers AT sourceware DOT cygnus DOT com
Message-ID: <36E837E4.58093232@cityweb.de>
Date: Thu, 11 Mar 1999 22:38:44 +0100
From: Corinna Vinschen <corinna DOT vinschen AT cityweb DOT de>
X-Mailer: Mozilla 4.04 [en] (WinNT; I)
MIME-Version: 1.0
To: Chris Faylor <cgf AT cygnus DOT com>, Earnie Boyd <earnie_boyd AT yahoo DOT com>
CC: cygwin-developers AT sourceware DOT cygnus DOT com
Subject: Re: Repost: process table shows already killed processes
References: <36E774FB DOT E74790F3 AT cityweb DOT de> <19990311111421 DOT A887 AT cygnus DOT com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Chris Faylor wrote:
> 
> I can't explain the behavior (which is why I haven't responded) but
> I can point out that if you are using some non-cygwin utility to kill
> the processes then the processes won't be able to clean themselves
> out of the cygwin process table.
> 
> cgf

Interesting point. I haven't the ultimate insight in the process
handling, so I wonder, that the cleanup seems to be done reproducable
on the second `ps' call.

Why am I forced, to use M$-kill instead of cygwin-kill? It's,
because I'm not able, to kill a process, which is started
as user `system' (from service manager via srvany) which
has no controlling terminal (setsid):
tcsh% kill -9 1001
1001: Not super-user.

I don't know, why this happens, especially, if I use my extension
for Unix-like permissions. If I use it, I'm able, to kill any
process of any user, but unfortunately not the above described ones.

Earnie Boyd wrote:
> Could the problem be a timing issue, where syslogd has more buffers to
> flush before disappearing or perhaps processing some event and
> captures the kill event before shutting down?  I notice that you've
> killed the processes in the reverse order that you started them, what
> happens if you kill syslogd first?

The order of killing is irrelevant. In any case, at least one of the
killed processes remains in the process table, until the second `ps'
call is done.

Regards,
Corinna