From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cuda.sgi.com (cuda1.sgi.com [192.48.157.11]) by oss.sgi.com (8.14.3/8.14.3/SuSE Linux 0.8) with ESMTP id p563lmGA200129 for ; Sun, 5 Jun 2011 22:47:48 -0500 Received: from mail.sandeen.net (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id 5FF041662030 for ; Sun, 5 Jun 2011 20:47:45 -0700 (PDT) Received: from mail.sandeen.net (sandeen.net [63.231.237.45]) by cuda.sgi.com with ESMTP id ZSVzicCGTgALii07 for ; Sun, 05 Jun 2011 20:47:45 -0700 (PDT) Message-ID: <4DEC4DE0.5090503@sandeen.net> Date: Sun, 05 Jun 2011 22:47:44 -0500 From: Eric Sandeen MIME-Version: 1.0 Subject: Re: altering defaults References: <8762onaq19.fsf@psinom.home> In-Reply-To: <8762onaq19.fsf@psinom.home> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: xfs-bounces@oss.sgi.com Errors-To: xfs-bounces@oss.sgi.com To: prad Cc: linux-xfs@oss.sgi.com On 6/3/11 10:33 AM, prad wrote: > i'm new to xfs (courtesy of most helpful and encouraging commentary by > stan hoeppner!) and i've seen some advice which says to make the > block size -> 512 > directory size -> 4096 > > on the other hand, i've also come across webpages which say don't mess > around! keep the defaults as they are unless you are absolutely sure > that changing it suits your purpose and know why. Yup. See "Q: I want to tune my XFS filesystems for " http://xfs.org/index.php/XFS_FAQ#Q:_I_want_to_tune_my_XFS_filesystems_for_.3Csomething.3E > my question is for the data storage area on a web/email server. we're > mainly going to have small files there and the email part will have only > temporary files for the most part since people will download (ie pop). > > it makes sense to make the block size = 512, but i wonder if it really > matters noticeably. the server is not a heavily visited one and only on > very rare occasions will we get around 50000 hits/day - most of the time > we're looking at under 10000 and that is with all the domains combined. If space efficiency is a consideration then smaller blocksizes might be useful. Not so much for hits per day, but for how many files you expect to store at any one time; you'll waste ((blocksize / 2) * nr files) on average. -Eric _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs