From owner-freebsd-chat Mon Dec 13 17: 2:44 1999 Delivered-To: freebsd-chat@freebsd.org Received: from mta3.rcsntx.swbell.net (mta3.rcsntx.swbell.net [151.164.30.27]) by hub.freebsd.org (Postfix) with ESMTP id BF4C51510C for ; Mon, 13 Dec 1999 17:02:41 -0800 (PST) (envelope-from noslenj@swbell.net) Received: from swbell.net ([207.193.44.221]) by mta3.rcsntx.swbell.net (Sun Internet Mail Server sims.3.5.1999.09.16.21.57.p8) with ESMTP id <0FMP008KOHK2N7@mta3.rcsntx.swbell.net> for chat@FreeBSD.ORG; Mon, 13 Dec 1999 19:02:31 -0600 (CST) Received: from localhost (localhost [127.0.0.1]) by swbell.net (8.9.3/8.9.3) with ESMTP id SAA00394; Mon, 13 Dec 1999 18:38:50 -0600 (CST envelope-from noslenj@swbell.net) Date: Mon, 13 Dec 1999 18:38:50 -0600 (CST) From: Jay Nelson Subject: Re: Log file systems? (Was: Re: dual 400 -> dual 600 worth it?) In-reply-to: <199912121124.MAA25648@zed.ludd.luth.se> To: Mattias Pantzare Cc: chat@FreeBSD.ORG Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-chat@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Thanks for the reply and the references. Your answer _should_ be a FAQ. It would save the repeat questions. -- Jay On Sun, 12 Dec 1999, Mattias Pantzare wrote: >> >How similar is that to the log partition in SGI's XFS? There was no >> >restriction as to what spindle the log filesystem was placed. Quite to >> >the contrary, it was indicated using a separate drive on a separate >> >SCSI bus would help performance. >> >> XFS sounds a lot like AIX's JFS. Which raises the question: What is >> the connection between BSD's lfs, soft updates, SGI's XFS and AIX's >> jfs? Don't they all do essentially the same thing except for where the >> log is written? > >No. lfs is a logging filesystem. You only have a log that contains everything, >including all your files. The downside to this is that you have to have a >garbage collector that cleans deleted data from the log. The good thing is >that you never have to seek for writes. All writes are to the end of the log. >http://www.cs.berkeley.edu/projects/sprite/papers/ has some papers on lfs. > >A journaling filesystem is like a normal filesystem but you have a transaction >log that turns synchronous writes into a synchronous write to the log and a >asynchronous write to the normal filesystem. This avoids seeks when latency is >important. > >Soft updates do not have a log att all. Take a look at > http://www.ece.cmu.edu/~ganger/papers/CSE-TR-254-95/. > > >Lfs will not roll in the normal sense, it will simply discard the half done >write at the end of the log if there is one. > >Soft updates can't do rolling as there is no log. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message