From owner-freebsd-fs Thu Oct 28 9:43:42 1999 Delivered-To: freebsd-fs@freebsd.org Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by hub.freebsd.org (Postfix) with ESMTP id 1F6CD14BC7 for ; Thu, 28 Oct 1999 09:43:35 -0700 (PDT) (envelope-from robert@cyrus.watson.org) Received: from fledge.watson.org (robert@fledge.pr.watson.org [192.0.2.3]) by fledge.watson.org (8.9.3/8.9.3) with SMTP id MAA31105; Thu, 28 Oct 1999 12:43:14 -0400 (EDT) (envelope-from robert@cyrus.watson.org) Date: Thu, 28 Oct 1999 12:43:14 -0400 (EDT) From: Robert Watson X-Sender: robert@fledge.watson.org Reply-To: Robert Watson To: Don Cc: freebsd-fs@freebsd.org Subject: Re: Journaling In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Wed, 27 Oct 1999, Don wrote: > Here is the question plain and simple: > Does anyone here want a journaled file system for FreeBSD? yes. > Such a file system would be built from the ground up to include such > things as the ability to grow and shrink the fs size, acl's on files, an > expandable architecture, journaling, etc. > > If people want it I will go through the effort of getting together the > programming talent necessary to do it. (Read: not me as I do not have > the skills to do anything other than manage a project like this) > > If people do not want it then I will drop it now. I think that having a journaled file system with extensible file attributes is a good idea--look at how easily XFS can be extended to support new features, yet to do this in a transactional way with meta-data. There are numerous concerns with adding new features to an FS, and having a journaled extensible FS would throw a bunch of them out the window (updating fsck to handle the new features, dealing with the consistency and atomicity issues, etc). That said, writing a journaled fs is not easy :-). BTW, I'd like to know how the layered file system people planned to address consistency between layers in the event of a failure -- i.e., a system crash occurs, how do you verify that the ACLs in an ACLfs layer are consistent with the attributes in the Attributefs layer, and that those are consistent with the actual files stored in the bottom layer? The only way to do this properly would seem to be through journaling and some kind of transaction system that knows about layers and treats it as a nested transaction? Robert N M Watson robert@fledge.watson.org http://www.watson.org/~robert/ PGP key fingerprint: AF B5 5F FF A6 4A 79 37 ED 5F 55 E9 58 04 6A B1 TIS Labs at Network Associates, Safeport Network Services To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message