From owner-freebsd-questions Tue Dec 11 4: 6: 4 2001 Delivered-To: freebsd-questions@freebsd.org Received: from tharmas.rintrah.org (dhcp065-024-235-184.insight.rr.com [65.24.235.184]) by hub.freebsd.org (Postfix) with SMTP id 1222737B405 for ; Tue, 11 Dec 2001 04:06:00 -0800 (PST) Received: (qmail 57615 invoked by uid 1000); 11 Dec 2001 12:06:17 -0000 Date: Tue, 11 Dec 2001 07:06:17 -0500 From: devin-freebsdquestions@rintrah.org To: Mark Hughes Cc: freebsd-questions@freebsd.org Subject: Re: Softupdates & fsck Message-ID: <20011211070617.A51634@tharmas.rintrah.org> Mail-Followup-To: Mark Hughes , freebsd-questions@freebsd.org References: <019301c1820a$1b07e2b0$0200a8c0@mark2> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <019301c1820a$1b07e2b0$0200a8c0@mark2>; from mh_lists@digitalspy.co.uk on Tue, Dec 11, 2001 at 06:07:19AM -0000 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Tue, Dec 11, 2001 at 06:07:19AM -0000, Mark Hughes wrote: > Guys, > > We've got softupdates enabled for all the filesystems, aside from /, on a > web server. Recently a power outage at our ISP (due to a blow fuse or some > such event) caused the power to our server to be cut completely for a brief > period, obviously without warning or a graceful shutdown. > > Now, when we brought it back up, the filesystem was in a little bit of a > mess, and we had to go into single user mode and use fsck to fix a number > of errors (log below). I was somewhat suprised at this, as I was under the > impression that softupdates whole purpose, aside from speed increases, was > to never leave the file system in a state where it would need an fsck if > the power was cut or the system crashed.... > > Was I wrong in this assumption? Or is something strange going on... > > Any ideas as to what has happened gratefully received. > > ad0s1a: hard error reading fsbn 656735 of 328304-328415 (ad0s1 bn 656735; > cn 40 tn 224 sn 23) status=59 error=40 > > > > CANNOT READ: BLK 328304 > > CONTINUE? [yn] y This was on / without softupdates. The worst of your errors. [snip] > > ** /dev/ad0s1h [snip] > UNREF FILE I=3166465 OWNER=nobody MODE=100644 > > SIZE=1295 MTIME=Dec 6 12:20 2001 > > RECONNECT? [yn] y lots of unreferenced files. You'll see that on the softupdates mounted partitions, fsck is only reclaiming unused space by clearing up these unreferenced files. [snip] > ** Phase 5 - Check Cyl groups > > FREE BLK COUNT(S) WRONG IN SUPERBLK > > SALVAGE? [yn] y > > > > SUMMARY INFORMATION BAD > > SALVAGE? [yn] y And here you have it. The summary info was bad. It's cleaning up the free blocks incorrectly marked used. [snip] The rest of your file systems looked pretty much the same to me. Didn't look like they were a "mess." Just reclaiming unused blocks as is supposed to happen with softupdates. As others have noted, you can do this in the background in -CURRENT. --devin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message