Goetz,

You might want to review your Posix man pages.  There is a kill -9.  I
even tried it on this stubborn session, with no success.  I'm not sure
that HP put the complete functionality into the kill command, but it is
there, and under normal conditions it works. (I tested this out).

I understand the stability issues, but this process was not doing
anything that should have the ability to corrupt anything.  I was using
sendmail interactively, pointing to a file for the input, and my NT box
crashed.  When I got back up and running, it had hosed the session to
the point that I could not abort it, the I/O on the ldev, or anything
else.  For some reason, the NS/VT session hung on.  After that, I got
creative in trying to kill this session.  I even tried the posix kill
-3, -6, and finally -9.  Nothing was working.  I even tried to chop some
of the files that it said was open, out from underneath it.  (Copy the
file somewhere else, and do a posix rm)  The file went away, but the
information on the process said that it was still using it.  (btw, copy
the file back so the next process can use it)  It is a little
frustrating when these don't go away, even after a network stop and
restart.

As a note, when using the kill -9 command, I tried to kill the youngest
process (the child of the child of the child.....) first, then I moved
up the chain until I got the first child process off of the system level
processes.  When I tried to do a kill on this, I was informed that I
could not perform the kill.  I figured that this was the HP3000's way of
guaranteeing stability.

Kevin Newman



> -----Original Message-----
> From: Götz Neumann [SMTP:[log in to unmask]]
> Sent: Thursday, April 29, 1999 1:16 AM
> To:   [log in to unmask]
> Subject:      Re: How do I kill a hung session ?
> 
> [log in to unmask] wrote:
> >
> > Re: Re: How do I kill a hung session ?
> >
> > Can't wait for ":abortproc" ?
> >
> > I usually try to stay away from the argumentative topics on the
> list, but
> > this one pulled me in.
> 
> Meeto, but I will share the flames with James :-)
> 
> Honestly, I am really proud of MPE not having a 'kill -9' equivalent.
> And of course I am the opposite of proud, whenever a new code defect
> leads to hung (unabortable) sessions or jobs, or 'ghost' or 'orphaned'
> processes.
>