From owner-freebsd-performance@FreeBSD.ORG Tue Feb 22 09:45:36 2005 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D29DD16A4CE; Tue, 22 Feb 2005 09:45:36 +0000 (GMT) Received: from av8-2-sn3.vrr.skanova.net (av8-2-sn3.vrr.skanova.net [81.228.9.184]) by mx1.FreeBSD.org (Postfix) with ESMTP id B54EC43D58; Tue, 22 Feb 2005 09:45:35 +0000 (GMT) (envelope-from martin@gneto.com) Received: by av8-2-sn3.vrr.skanova.net (Postfix, from userid 502) id 669F137F1A; Tue, 22 Feb 2005 10:45:34 +0100 (CET) Received: from smtp3-2-sn3.vrr.skanova.net (smtp3-2-sn3.vrr.skanova.net [81.228.9.102]) by av8-2-sn3.vrr.skanova.net (Postfix) with ESMTP id 4FD7B37EA9; Tue, 22 Feb 2005 10:45:34 +0100 (CET) Received: from [192.168.2.30] (h118n1fls31o985.telia.com [213.65.16.118]) by smtp3-2-sn3.vrr.skanova.net (Postfix) with ESMTP id 093FE37E47; Tue, 22 Feb 2005 10:45:33 +0100 (CET) Message-ID: <421AFF3C.8090608@gneto.com> Date: Tue, 22 Feb 2005 10:45:32 +0100 From: Martin Nilsson User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206) X-Accept-Language: sv, en-us, en MIME-Version: 1.0 To: Robert Watson References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit cc: David Rice cc: freebsd-performance@FreeBSD.org Subject: Re: High traffic NFS performance and availability problems X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 22 Feb 2005 09:45:36 -0000 Robert Watson wrote: > In the top output below, it looks like there's a lot of contention on > Giant. In 5.3-RELEASE and before, the amr driver is not MPSAFE, but my > understanding is that in 5-STABLE, it has been made MPSAFE, which may make > quite a difference in performance. I pinged Scott Long, who did the work The locking of amr have not been MFC:d to RELENG_5 yet, I've just checked with webcvs. From my brief and unconclusive tests the locked driver was much faster (about 2x on sequential transfers). /Martin