From owner-freebsd-fs Tue Oct 24 11:48:29 1995 Return-Path: owner-fs Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id LAA25941 for fs-outgoing; Tue, 24 Oct 1995 11:48:29 -0700 Received: from aslan.cdrom.com (aslan.cdrom.com [192.216.223.142]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id LAA25936 for ; Tue, 24 Oct 1995 11:48:27 -0700 Received: from localhost.cdrom.com (localhost.cdrom.com [127.0.0.1]) by aslan.cdrom.com (8.6.12/8.6.9) with SMTP id LAA24458; Tue, 24 Oct 1995 11:47:40 -0700 Message-Id: <199510241847.LAA24458@aslan.cdrom.com> X-Authentication-Warning: aslan.cdrom.com: Host localhost.cdrom.com didn't use HELO protocol To: Tom Pavel cc: "Justin T. Gibbs" , Peter Dufault , freebsd-fs@FreeBSD.org Subject: Re: WORM fs In-reply-to: Your message of "Tue, 24 Oct 1995 11:35:23 PDT." <199510241835.LAA16495@MAILBOX.SLAC.Stanford.EDU> Date: Tue, 24 Oct 1995 11:47:40 -0700 From: "Justin T. Gibbs" Sender: owner-fs@FreeBSD.org Precedence: bulk > >>>>>> On Mon, 23 Oct 1995, "Justin T. Gibbs" >write >s: > >> >Are any of the working file systems appropriate for a WORM drive >> >for an archival application? >> > >> >I would like to be able to mount the file system and search through >> >the files in the usual manner. It can be unmounted for writing. >> >> LFS would suit this task well, but it doesn't work in -current. > > >I'm not sure what the problems with LFS in FreeBSD-current are, but just >for general info in case people weren't aware, there was a lot of work on >LFS for BSD4.4 done by Margo Seltzer and co. You can find a discussion of >various related issues in: > http://das-www.harvard.edu/users/faculty/Margo_Seltzer/usenix.195.html >and you can get the mods to the LFS sources in : > ftp://virtual.harvard.edu/pub/margo/usenix.195/lfs.tar.gz >The papers are interesting reading, as is Ousterhout's response. Yup. I've had these patches sitting on my shelf since late December of last year. Mostly they allow LFS to use frags but there are some bug fixes in there too. The reason that LFS doesn't work in -current is that the whole implementation breaks VFS rules left and right, and when John changed some of the VFS interfaces for better performance a while back, LFS broke. I encourange anyone who has time to look into LFS, but be forwarned that portions of it are ugly as sin and it needs some large architectural changes to make it live up to its performance expectations. >I haven't actually checked, but my understanding is that these mods have >not been worked into either the FreeBSD or NetBSD kernels. If that is >correct, then they would probably provide a good starting point for someone >to get LFS working. NetBSD should be able to bring those changes in fairly easily. I need to first get the old code working under -current then merge all of my bug fixes in with Seltzer's code. >For what it's worth, It could be worth a lot. The benchmarks I got last year using LFS were quite impressive and this is without the many optimizations that are apprant when you wade throught the code. >Tom Pavel > >Stanford Linear Accelerator Center >pavel@slac.stanford.edu http://www.slac.stanford.edu/~pavel/ -- Justin T. Gibbs =========================================== Software Developer - Walnut Creek CDROM FreeBSD: Turning PCs into workstations ===========================================