From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: with ECARTIS (v1.0.0; list xfs); Sun, 18 May 2008 06:18:11 -0700 (PDT) Received: from cuda.sgi.com (cuda2.sgi.com [192.48.168.29]) by oss.sgi.com (8.12.11.20060308/8.12.11/SuSE Linux 0.7) with ESMTP id m4IDI7nh031853 for ; Sun, 18 May 2008 06:18:09 -0700 Received: from bombadil.infradead.org (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id 75F07184B67 for ; Sun, 18 May 2008 06:18:56 -0700 (PDT) Received: from bombadil.infradead.org (bombadil.infradead.org [18.85.46.34]) by cuda.sgi.com with ESMTP id 4FOxnIFsWmOMuf4s for ; Sun, 18 May 2008 06:18:56 -0700 (PDT) Date: Sun, 18 May 2008 09:18:55 -0400 From: Christoph Hellwig Subject: Re: XFS filesystem reports as full though it isn't Message-ID: <20080518131855.GA1383@infradead.org> References: <20080516222755.3e557c00@galadriel.home> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20080516222755.3e557c00@galadriel.home> Sender: xfs-bounce@oss.sgi.com Errors-to: xfs-bounce@oss.sgi.com List-Id: xfs To: Emmanuel Florac Cc: xfs@oss.sgi.com On Fri, May 16, 2008 at 10:27:55PM +0200, Emmanuel Florac wrote: > cluster1:/proc/sys/fs# xfs_db -r -c 'freesp -s -a 0' /dev/vg0/lv0 > from to extents blocks pct > 1 1 5 5 71.43 > 2 3 1 2 28.57 > total free extents 6 > total free blocks 7 > average free extent size 1.16667 > > 7 blocks free? But there are 9 TB of free space! The -a 0 means you're only looking at the first AG, which given the history of the filesystem has a fair chance to be full. To get the summary for all AG do a xfs_db -r -c 'freesp -s' /dev/vg0/lv0 No idea yet on what could cause your problem, sorry. > The only significant thing I see is that the FS is really close to 16 > TB of allocated data (15.7TB). I tried mounting it with "inode64" > option with no more loving. Did you remount it with that option as in mount -o remount or did you completely unmount it and mount it again? Unfortunately none of the xfs-specific mount options can be reset with remount yet, and it doesn't return an error either (both are added to my todo list now)