From owner-freebsd-fs@FreeBSD.ORG Fri Mar 4 13:41:18 2011 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 1A9901065673 for ; Fri, 4 Mar 2011 13:41:18 +0000 (UTC) (envelope-from mcdouga9@egr.msu.edu) Received: from mail.egr.msu.edu (gribble.egr.msu.edu [35.9.37.169]) by mx1.freebsd.org (Postfix) with ESMTP id E3A348FC17 for ; Fri, 4 Mar 2011 13:41:17 +0000 (UTC) Received: from gribble (localhost [127.0.0.1]) by mail.egr.msu.edu (Postfix) with ESMTP id 9E1FD6E9F2 for ; Fri, 4 Mar 2011 08:23:29 -0500 (EST) X-Virus-Scanned: amavisd-new at egr.msu.edu Received: from mail.egr.msu.edu ([127.0.0.1]) by gribble (gribble.egr.msu.edu [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Bl83YsjmY0JX for ; Fri, 4 Mar 2011 08:23:29 -0500 (EST) Received: from [35.9.44.65] (daemon.egr.msu.edu [35.9.44.65]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (No client certificate requested) (Authenticated sender: mcdouga9) by mail.egr.msu.edu (Postfix) with ESMTPSA id 7A78C6E9BA for ; Fri, 4 Mar 2011 08:23:29 -0500 (EST) Message-ID: <4D70E7D1.5030205@egr.msu.edu> Date: Fri, 04 Mar 2011 08:23:29 -0500 From: Adam McDougall User-Agent: Mozilla/5.0 (X11; U; FreeBSD amd64; en-US; rv:1.9.2.13) Gecko/20110101 Thunderbird/3.1.7 MIME-Version: 1.0 To: freebsd-fs@freebsd.org References: <1299232133.18671.3.camel@pc286.embl.fr> In-Reply-To: <1299232133.18671.3.camel@pc286.embl.fr> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit Subject: Re: kmem_map too small with ZFS and 8.2-RELEASE X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 04 Mar 2011 13:41:18 -0000 On 03/04/11 04:48, Mickaël Canévet wrote: >> I'd use vm.kmem_size="32G" (i.e. twice your RAM) and that's it. > > Should I also increase vfs.zfs.arc_max ? > > Do you have any idea why the kernel panicked at only 8GB allocated ? > > Thank you I believe ARC allocations in kmem can become fragmented, so when it is searching for a place to store a new contiguous segment of memory, the remaining fragmented free spaces may all be too small. I also set vm.kmem_size to about twice the amount of ram to help it avoid this issue. I suspect if kmem is badly fragmented then performance of ZFS can downgrade, so that is another reason to keep kmem bigger.