From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753305Ab1CADvH (ORCPT ); Mon, 28 Feb 2011 22:51:07 -0500 Received: from ipmail06.adl6.internode.on.net ([150.101.137.145]:50473 "EHLO ipmail06.adl6.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752835Ab1CADvG (ORCPT ); Mon, 28 Feb 2011 22:51:06 -0500 X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AvsEAC78a015LFEb/2dsb2JhbACmSnW7RA2FVAQ Date: Tue, 1 Mar 2011 14:51:01 +1100 From: Dave Chinner To: Otto Meier Cc: linux-kernel@vger.kernel.org Subject: Re: [regression]2.6.38-rc6 with XFS still does consume all free space Message-ID: <20110301035101.GH3087@dastard> References: <4D63A9E7.3000803@gmx.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4D63A9E7.3000803@gmx.net> User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Feb 22, 2011 at 01:19:51PM +0100, Otto Meier wrote: > Hello, > on my System a "make modules_install" with kernel 2.6.38-rc6 has eaten up > all free Space on a XFS partition. Are you building the 2.6.38-rc6 kernel on some other kernel or are you running 2.6..38-rc6? > The free Space is > > /dev/sda6 56472984 43613692 12859292 78% / > > after make modules_install i get > > /dev/sda6 56472984 56472964 20 100% / Had you already built the kernel, or did the make command build the kernel and all the modules before installing them? Can you check how much space the modules directory is actually using (i.e. using du)? Also, did you get an ENOSPC error? > with kernel 2.6.37.1 everything runs fine and df shows > > /dev/sda6 56472984 43262256 13210728 77% / > > any idea what is going on here? No, all the known preallocation regressions have been fixed. can you confirm that the kernel you are running has this commit in it? b8fc826 xfs: speculative delayed allocation uses rounddown_power_of_2 badly Cheers, Dave. -- Dave Chinner david@fromorbit.com