From owner-freebsd-fs@FreeBSD.ORG Sat Jul 5 14:05:38 2014 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 9FF1E855; Sat, 5 Jul 2014 14:05:38 +0000 (UTC) Received: from systemdatarecorder.org (ec2-54-246-96-61.eu-west-1.compute.amazonaws.com [54.246.96.61]) (using TLSv1.1 with cipher ECDHE-RSA-AES256-SHA (256/256 bits)) (Client CN "localhost", Issuer "localhost" (not verified)) by mx1.freebsd.org (Postfix) with ESMTPS id 2FEED2D1B; Sat, 5 Jul 2014 14:05:36 +0000 (UTC) Received: from nereid (84-253-211-213.bb.dnainternet.fi [84.253.211.213]) (authenticated bits=0) by systemdatarecorder.org (8.14.4/8.14.4/Debian-2ubuntu2.1) with ESMTP id s65E3ugG030946 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Sat, 5 Jul 2014 14:03:57 GMT Date: Sat, 5 Jul 2014 17:05:24 +0300 From: Stefan Parvu To: Roger Pau =?ISO-8859-1?Q?Monn=E9?= Subject: Re: Strange IO performance with UFS Message-Id: <20140705170524.4212b6fa0b1046a33e1fc69a@systemdatarecorder.org> In-Reply-To: <53B7C616.1000702@citrix.com> References: <53B691EA.3070108@citrix.com> <53B69C73.7090806@citrix.com> <20140705001938.54a3873dd698080d93d840e2@systemdatarecorder.org> <53B7C616.1000702@citrix.com> Organization: systemdatarecorder.org X-Mailer: Sylpheed 3.4.1 (GTK+ 2.24.22; amd64-portbld-freebsd11.0) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: freebsd-fs@freebsd.org, FreeBSD Hackers X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 05 Jul 2014 14:05:38 -0000 > This looks much better than what I've saw in my benchmarks, how much > memory does the system have? We use on this system 64GB RAM. If you increase the block size in fio you should see better throughput, as you already found. Cool, you sorted out the thing. As a side note: interesting for us, was to discover that system usage between Debian 7 and FreeBSD was kind of different for our test workloads. Strange Linux system was around 3-4% system time, no matter what sort of block size or number of files we were pushing using hardware raid 10, resulting in a high iowait time and high run queue length (which on Linux systems adds to it the iowait). FreeBSD, I think, does not add to the run queue length the iowait processes waiting for a storage, network etc. Is this correct ? -- Stefan Parvu