From owner-freebsd-fs@FreeBSD.ORG Tue Dec 15 22:17:29 2009 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id BFC15106566C for ; Tue, 15 Dec 2009 22:17:29 +0000 (UTC) (envelope-from areilly@bigpond.net.au) Received: from nskntmtas03p.mx.bigpond.com (nskntmtas03p.mx.bigpond.com [61.9.168.143]) by mx1.freebsd.org (Postfix) with ESMTP id 495988FC1A for ; Tue, 15 Dec 2009 22:17:28 +0000 (UTC) Received: from nskntotgx02p.mx.bigpond.com ([124.188.161.100]) by nskntmtas03p.mx.bigpond.com with ESMTP id <20091215221727.XLDM1310.nskntmtas03p.mx.bigpond.com@nskntotgx02p.mx.bigpond.com>; Tue, 15 Dec 2009 22:17:27 +0000 Received: from duncan.reilly.home ([124.188.161.100]) by nskntotgx02p.mx.bigpond.com with ESMTP id <20091215221727.YNUF5306.nskntotgx02p.mx.bigpond.com@duncan.reilly.home>; Tue, 15 Dec 2009 22:17:27 +0000 Date: Wed, 16 Dec 2009 09:17:27 +1100 From: Andrew Reilly To: Hywel Mallett Message-ID: <20091215221727.GA8137@duncan.reilly.home> References: <20091208224710.GA97620@duncan.reilly.home> <228D9370-4967-4C47-9746-8475DCD4FA27@hmallett.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <228D9370-4967-4C47-9746-8475DCD4FA27@hmallett.co.uk> User-Agent: Mutt/1.4.2.3i X-Authentication-Info: Submitted using SMTP AUTH LOGIN at nskntotgx02p.mx.bigpond.com from [124.188.161.100] using ID areilly@bigpond.net.au at Tue, 15 Dec 2009 22:17:27 +0000 X-RPD-ScanID: Class unknown; VirusThreatLevel unknown, RefID str=0001.0A150204.4B280AF7.00A9,ss=1,fgs=0 X-SIH-MSG-ID: rBg1Edb/TAD0zmQs0WyzOwJxyArnqyN48Z4QX81loRIGTUDCp8DeQ9rHNvZRtdu1xD9JJhiHNGAnaa7jTY3RstCK Cc: freebsd-fs@freebsd.org Subject: Re: On gjournal vs unexpected shutdown (-->fsck) X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Dec 2009 22:17:29 -0000 On Tue, Dec 15, 2009 at 09:49:56PM +0000, Hywel Mallett wrote: > > On 8 Dec 2009, at 22:47, Andrew Reilly wrote: > > > Hi there, > > > > I thought that I'd try a gjournal'd UFS on one of my spare > > drives (so: dedicated to the task, formatted from clean, per the > > instructions in the gjournal man page.) The filesystem itself > > seems to be working swimmingly, although it isn't heavily used. > > In the time that I've had it running, though, I've had two power > > outages that have resulted in unexpected shutdowns, and I was > > surprised to find that the boot process did nothing unexpected: > > file system not marked clean: fsck before you can mount. So > > both times I fsck'd the drive, and as near as I can tell this > > took exactly as long as fsck on a regular UFS system of similar > > size. Isn't the journalling operation supposed to confer a > > shortcut benefit here? I know that the man page doesn't mention > > recovery by journal play-back, but I thought that it didn't need > > to: that's the whole point. Is there a step that I'm missing? > > Perhaps a gjournal-aware version of fsck that I should run > > instead of regular fsck, that will quickly mark the file system > > clean? > > > > (Running -current as of last weekend, if that matters.) > > > I assume you've run tunefs -J enable on the filesystems on > the journalled provider? Or used newfs with -J if it's a new > filesystem? > > If I remember correctly it's this flag that fsck checks to see > whether fsck is needed or not. > > You can check whether the flag is set or not by running dumpfs > on the filesystem. Under "flags" it'll say gjournal if the > flag is set. I've just taken the file system off-line and run tunefs -J enable on it, and tunefs said: tunefs: gjournal remains unchanged as enabled so I seem to have set it up properly in the first place. In the "further reading" list on the gjournal article, there is mention of mounting with async,gjournal options, but I see no reference to gjournal in the man pages, so my guess is that this is what has been superceded by the -J tunefs/newfs flag? Cheers, -- Andrew