From mboxrd@z Thu Jan 1 00:00:00 1970 From: "girish" Subject: Re: [PATCH] 2.5.x write_super is not for syncing Date: Thu, 5 Dec 2002 22:54:15 +0530 Sender: linux-fsdevel-owner@vger.kernel.org Message-ID: <001901c29c83$2300a990$9e744c0f@nt16158> References: <20021204192358.T27991@parcelfarce.linux.theplanet.co.uk> <20021205103601.GA23972@conectiva.com.br> <1039105890.2728.17.camel@sisko.scot.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Cc: "Bryan Henderson" , "Matthew Wilcox" , "Andrew Morton" , , "Chris Mason" , "Stephen Tweedie" Return-path: To: "Stephen C. Tweedie" , "Arnaldo Carvalho de Melo" List-Id: linux-fsdevel.vger.kernel.org > Hi, > > On Thu, 2002-12-05 at 10:36, Arnaldo Carvalho de Melo wrote: > > Em Wed, Dec 04, 2002 at 02:17:40PM -0800, Bryan Henderson escreveu: > > > > It's Storage Tank File System -- a shared filesystem type that IBM has been > > > talking about making for a couple of years. It's supposed to allow you to > > > put a single filesystem on a large number of shared disks and have all the > > > processes on all those systems access the filesystem as if they were all on > > > one big system. But those disks contain only data blocks. Metadata is > > > kept behind a server which is accessed via conventional network interfaces. > > > Isn't this what Peter Braam's Lustre FS is doing? http://www.lustre.org > > Yes, sounds very similar indeed. In lustre we have network connection between client and data servers, where as STFS is similar to CXFS from SGI Regards Girish