From owner-freebsd-current  Tue Sep 10 13:20:05 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id NAA10886
          for current-outgoing; Tue, 10 Sep 1996 13:20:05 -0700 (PDT)
Received: from spinner.DIALix.COM (spinner.DIALix.COM [192.203.228.67])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id NAA10821;
          Tue, 10 Sep 1996 13:19:57 -0700 (PDT)
Received: from spinner.DIALix.COM (localhost.DIALix.oz.au [127.0.0.1])
          by spinner.DIALix.COM (8.7.5/8.7.3) with ESMTP id EAA00788;
          Wed, 11 Sep 1996 04:19:45 +0800 (WST)
Message-Id: <199609102019.EAA00788@spinner.DIALix.COM>
X-Mailer: exmh version 1.6.7 5/3/96
To: "Jordan K. Hubbard" <jkh@time.cdrom.com>
cc: dyson@freebsd.org, phk@critter.tfs.com (Poul-Henning Kamp),
        current@freebsd.org
Subject: Re: VM/kernel problems? 
In-reply-to: Your message of "Tue, 10 Sep 1996 10:54:00 MST."
             <20530.842378040@time.cdrom.com> 
Date: Wed, 11 Sep 1996 04:19:45 +0800
From: Peter Wemm <peter@spinner.DIALix.COM>
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

"Jordan K. Hubbard" wrote:
> > Okay, I admit it... It's broke... :-).  If it isn't working tonight,
> > it will be tomorrow by backing out the errant changes.
> 
> One possible wrinkle: I'm not seeing this at all with this morning's
> kernel but I have 96MB of main memory.  Perhaps this is related to
> reclaimation in low memory situations?  The processes getting shot do
> seem to be the larger ones, predominantly.
> 
> 						Jordan

Also, this is subjective, but the machine seems to feel less responsive as 
time progresses.  This one has 48M and does little more than serve NFS, 
run X, compiles, and a zillion xterm's, tcsh and ssh sessions.

I recompiled the kernel making a "try this if you get the chance" change 
from John, and the problem has not yet surfaced after a few hours.  On the 
other hand, it hasn't felt like it's slowing down either (yet?), so I 
don't know if it's hard-up for memory yet.  It's too soon to tell for me, 
I think.

The other machines are all 16M, do occasional massive NFS client attacks 
on the above machine (cvs update over NFS is a killer) and lots of 
recompiles.  The problem seemed to be striking most then for processes 
that had been paged out, but that was not consistant. named died when the 
machine had been idle while I was asleep.

Cheers,
-Peter