From mboxrd@z Thu Jan 1 00:00:00 1970 From: Milind Subject: Re: EXT vs XFS at 80% filled filesystem Date: Mon, 18 May 2009 18:33:28 +0530 Message-ID: <1242651808.3339.28.camel@alhena> References: <49F9565E.40804@gslab.com> <20090430183450.GB19276@mit.edu> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: linux-fsdevel@vger.kernel.org, xfs@oss.sgi.com To: Theodore Tso Return-path: Received: from padma.gslab.com ([59.163.66.102]:60136 "HELO padma.gslab.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751794AbZERNDd (ORCPT ); Mon, 18 May 2009 09:03:33 -0400 In-Reply-To: <20090430183450.GB19276@mit.edu> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Hi Theodore, I am facing some weird problem of cross compiling libuuid sources. I downloaded e2fsprogs and cross-compiled them but when I do "make install-libs" it doesn't really install libuuid.so but installs libuuid.a. Can I have some pointers to build libuuid.so to have it in my toolchain libraries. I have crawled through web but couldn't get any nice solution to this. Thanks in Advance -Miline On Thu, 2009-04-30 at 14:34 -0400, Theodore Tso wrote: > On Thu, Apr 30, 2009 at 01:12:22PM +0530, Milind Dumbare wrote: > > Hi, > > > > I have heard of XFS's performance is not good as compared to EXT3 when > > the filesystem(disk) is 80% filled with data. Is it true? I have went > > through lots of performance documents of both XFS and EXT3 but could not > > find such performance benchmarking (for 80% full filesystems). > > I've not heard of any such performance metrics, and I suspect it would > very much depend on how the filesystem was "aged". A filesystem that > has been in use for a few years and is at 80% capacity will behave > very different from a brand-new filesystem which was freshly formatted > and then filled with a few large files until said filesystem was 80% > full. > > - Ted >