From owner-freebsd-stable@FreeBSD.ORG Wed Mar 5 21:18:06 2008 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C9A9D1065671 for ; Wed, 5 Mar 2008 21:18:06 +0000 (UTC) (envelope-from kris@FreeBSD.org) Received: from weak.local (freefall.freebsd.org [IPv6:2001:4f8:fff6::28]) by mx1.freebsd.org (Postfix) with ESMTP id D803C8FC1B; Wed, 5 Mar 2008 21:18:05 +0000 (UTC) (envelope-from kris@FreeBSD.org) Message-ID: <47CF0E0C.70001@FreeBSD.org> Date: Wed, 05 Mar 2008 22:18:04 +0100 From: Kris Kennaway User-Agent: Thunderbird 2.0.0.12 (Macintosh/20080213) MIME-Version: 1.0 To: Michael Grant References: <62b856460803050106t4375c59aj586a8b48bf91f7de@mail.gmail.com> <62b856460803051029v71f4adacia5f43cd3ecb66db3@mail.gmail.com> In-Reply-To: <62b856460803051029v71f4adacia5f43cd3ecb66db3@mail.gmail.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-stable@freebsd.org Subject: Re: swap_pager: indefinite wait buffer 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, 05 Mar 2008 21:18:06 -0000 Michael Grant wrote: > On Wed, Mar 5, 2008 at 11:08 AM, Ruben van Staveren wrote: >> On 5 Mar 2008, at 10:06, Michael Grant wrote: >> >> > My server just literally was brought to it's knees with this message >> > spewing on the console: >> > >> > swap_pager: indefinite wait buffer: bufobj: 0, blkno: 1203133, size: >> > 4096 >> > >> > (blkno and size were varying) >> > >> > Some searching says that this is or was a bug. Has this been fixed >> > yet? If so, what should I upgrade to? I'm currently running 6.3 >> >> You may consider partition backed swap instead of file backed swap if >> that is the case. > > Hmm, I can't easily do that, I didn't leave any empty partitions > around as I never considered swapping to a file to be a so bad. > > Is swapping to a file so bad under normal conditions? The message indicates that it took >30 seconds to complete an operation, so it was timed out assuming the I/O was lost by the device. In your case it was probably not lost, just delayed for more than 30 seconds by an overloaded filesystem. > Does this mean that this bug is still not fixed in 7.0? It's not clear whether it's a bug or your disk is just too overloaded to complete the filesystem operation in a reasonable time period (swapping to a file is slower than swapping to a partition, which is already something you never want to do in normal operation). You can increase the timeout by editing the kernel. Kris