Date: Fri, 30 Jan 1998 21:26:36 -0600 From: Jonathan Lemon <jlemon@americantv.com> To: Ollivier Robert <roberto@keltia.freenix.fr> Cc: cvs-committers@FreeBSD.ORG Subject: Re: cvs commit: src/sys/conf files options src/sys/i386/conf LINT src/sys/i386/i386 autoconf.c src/sys/kern init_main.c init_sys Message-ID: <19980130212636.64874@right.PCS> In-Reply-To: <19980131030605.34628@keltia.freenix.fr>; from Ollivier Robert on Jan 01, 1998 at 03:06:05AM %2B0100 References: <015201bd2dc9$c7e51f00$2844c00a@cello.synapse.net> <19980130162329.10474@right.PCS> <19980131030605.34628@keltia.freenix.fr>
next in thread | previous in thread | raw e-mail | index | archive | help
On Jan 01, 1998 at 03:06:05AM +0100, Ollivier Robert wrote: > According to Jonathan Lemon: > > Well, from what I understand of LFS, it performs best as a 'write-only' > > filesystem. You aren't really supposed to be reading from an LFS, you > > should be reading from your memory cache. From that perspective, I heard > > one professor call a 64MB cache "small" for a LFS system. > > If you read Margo's thesis, the huge memory consumption is the big problem > of Sprite and LFS is supposed to need less memory. > > One of its biggest problem apart from being incompatible with the unified > buffer cache/VM is that many features were never implemented (like the > rollback fsck). > > The thesis is very enlightning on FS and their flaws BTW. Reference? I think I only have papers from Rosenblum and Ousterhout. -- Jonathan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980130212636.64874>