From owner-freebsd-stable@FreeBSD.ORG Tue Jul 29 06:55:21 2003 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 70C0637B401 for ; Tue, 29 Jul 2003 06:55:21 -0700 (PDT) Received: from phys9911.phys.tue.nl (PHYS9911.phys.tue.nl [131.155.108.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5768B43FA3 for ; Tue, 29 Jul 2003 06:55:20 -0700 (PDT) (envelope-from karelj@phys9911.phys.tue.nl) Received: from phys9911.phys.tue.nl (localhost [127.0.0.1]) by phys9911.phys.tue.nl (8.12.9/8.12.9) with ESMTP id h6TDtEfF011044; Tue, 29 Jul 2003 15:55:14 +0200 (CEST) (envelope-from karelj@phys9911.phys.tue.nl) Received: (from karelj@localhost) by phys9911.phys.tue.nl (8.12.9/8.12.9/Submit) id h6TDtBSK011043; Tue, 29 Jul 2003 15:55:11 +0200 (CEST) Date: Tue, 29 Jul 2003 15:55:11 +0200 From: "Karel J. Bosschaart" To: Garance A Drosihn Message-ID: <20030729135511.GA10921@phys9911.phys.tue.nl> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.1i cc: stable@freebsd.org cc: Kirk McKusick Subject: Re: Strange results after partition-full condition... X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 29 Jul 2003 13:55:21 -0000 On Tue, Jul 29, 2003 at 12:55:26AM +0200, Garance A Drosihn wrote: > I have a test 4.x-stable system that I recently rebooted into. > The last time I had updated it was May 3rd. I cvsup'ed it, > did the buildworld/installworlds, and everything seemed fine. > > I then thought I would update all the ports. When upgrading > XFree86, the /usr partition ran out of disk space. Now the > partition shows up as: > > (21) df -k > Filesystem 1K-blocks Used Avail Capacity Mounted on > /dev/ad0s2a 251950 35550 196244 15% / > /dev/ad0s2f 241870 10 222512 0% /tmp > /dev/ad0s2g 2064302 -164180 2063338 -9% /usr > /dev/ad0s2e 257998 61258 176102 26% /var > /dev/ad0s2h 1311026 676644 529500 56% /Users > > This has persisted through a bunch of 'sync's and a system > reboot. Actually it had started as -1% capacity, but went > to -9% as I removed files (like all of /usr/obj/usr/src ). > A while back I've also seen weird numbers on one of my partitions: despite that minfree was 8% (default) the "Capacity" went to ~130%, and the "Used" part became larger than the number of 1K-blocks. This was on -stable, soft-updates enabled. 'sync's didn't have any effect either. > I have some other process running right now, but when that is > done I'm going to shutdown and then run fsck on that partition. > I assume that will clear it up. > Yes, in my case it solved the problem. > This is on a dual-CPU system, if that is significant. The > partition is mounted: > /dev/ad0s2g on /usr (ufs, local, soft-updates) > > This is only a test system, so it isn't much of a problem for > me. I just thought that it was odd enough that I should > mention it. Has anyone else seen behavior like this? > I've seen it, but no idea how to reproduce above situation. I don't know if it is related, but when using a USB flash drive (UFS formatted, no soft-updates, minfree=0), I often see after mounting that the free space is incorrect; it shows the free space it had before I deleted some files, during a previous mount. Unmounting and fsck'ing solves it. Usually it happens when the stick is unmounted in -stable and mounted on -current. Although mounting in -stable shows the correct numbers, an fsck (also on -stable) reveals "SUMMARY INFORMATION BAD". Is this an indication that the filesystem was left in a dirty state? Both -stable and -current mount it anyway, but only on -current it is evident from the numbers that something is wrong. (After unmounting I always wait some time before disconnecting the flash drive, at least until the LED doesn't flash anymore, but usually much longer). Karel.