[ On Monday, December 9, 2002 at 10:08:41 (-0500), Gary Thorpe wrote: ]
> Subject: Re: NetBSD1.6 UVM problem? 
>
> Its funny how people can rail about how a 0.5 sec delay on fork() when
> a process limit is reached is "punishing" processes unfairly but have
> no problem with killing them at random when memory is overcommited...

There have been several threads on this list (and others) where people
have railed on about the horrendousness of "random" process killing.

Indeed this debate started very long ago, long before NetBSD in fact.

My first complaint about it was sometime around 1990 or so when I had
AIX-3.2 kill my Xserver right out from under me because someone else on
the same system was trying to run some other big program.  Complaints
like mine (perhaps even before 3.2) are what caused IBM to eventually
introduce the hack of a catchable SIGDANGER at the first sign of the
ceiling coming close and then start killing those which ignored it once
the ceiling is really hit.

-- 
								Greg A. Woods

+1 416 218-0098;            <g.a.woods@ieee.org>;           <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>