From owner-freebsd-stable@FreeBSD.ORG Wed Aug 6 07:13:22 2008 Return-Path: Delivered-To: stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B8D21106566B for ; Wed, 6 Aug 2008 07:13:22 +0000 (UTC) (envelope-from ericlin@tamama.org) Received: from rv-out-0506.google.com (rv-out-0506.google.com [209.85.198.237]) by mx1.freebsd.org (Postfix) with ESMTP id 9973C8FC1B for ; Wed, 6 Aug 2008 07:13:22 +0000 (UTC) (envelope-from ericlin@tamama.org) Received: by rv-out-0506.google.com with SMTP id b25so4326234rvf.43 for ; Wed, 06 Aug 2008 00:13:22 -0700 (PDT) Received: by 10.140.144.6 with SMTP id r6mr8689550rvd.293.1218006802209; Wed, 06 Aug 2008 00:13:22 -0700 (PDT) Received: by 10.141.68.8 with HTTP; Wed, 6 Aug 2008 00:13:22 -0700 (PDT) Message-ID: <47713ee10808060013h10dd3f57ma5f45e69a322743a@mail.gmail.com> Date: Wed, 6 Aug 2008 15:13:22 +0800 From: "Lin Jui-Nan Eric" To: stable@freebsd.org In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <47713ee10808050839k5b258831x66bc52f70b2c355b@mail.gmail.com> Cc: Subject: Re: Max size of one swap slice X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 06 Aug 2008 07:13:22 -0000 On Wed, Aug 6, 2008 at 12:46 AM, Chuck Swiger wrote: > It's hard to conceive of why you'd want to add so much swap space, anyway-- > if you've got programs which actually need to deal with 10s of gigabytes > worth of data, then they ought to maintain a smaller/reasonable-sized > working set in RAM and do disk I/O as needed themselves rather than depend > upon the VM pager, anyways. We are running varnish, and found that it is not stable while using mmap mode. We don't know whether if the problem is in the code of varnish or file system, but we found that if we run varnish using malloc mode with big swap, it became stable. Thank you all for the information, I'll try to look into the kernel code.