From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S268890AbUHLXtJ (ORCPT ); Thu, 12 Aug 2004 19:49:09 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S268893AbUHLXtI (ORCPT ); Thu, 12 Aug 2004 19:49:08 -0400 Received: from thunk.org ([140.239.227.29]:16107 "EHLO thunker.thunk.org") by vger.kernel.org with ESMTP id S268890AbUHLXtF (ORCPT ); Thu, 12 Aug 2004 19:49:05 -0400 Date: Thu, 12 Aug 2004 19:48:57 -0400 From: "Theodore Ts'o" To: Andi Kleen Cc: linux-kernel@vger.kernel.org Subject: Re: New concept of ext3 disk checks Message-ID: <20040812234857.GA8098@thunk.org> Mail-Followup-To: Theodore Ts'o , Andi Kleen , linux-kernel@vger.kernel.org References: <2ssbz-jB-1@gated-at.bofh.it> <2swyz-3ny-13@gated-at.bofh.it> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.6+20040803i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Aug 13, 2004 at 01:24:33AM +0200, Andi Kleen wrote: > Theodore Ts'o writes: > > > 4) If there were no errors detecting by the fsck run, run the command > > "tune2fs -C 0 -T now /dev/XXX" on the live filesystem. This sets the > > mount count and last filesystem checked time to the appropriate values > > in the superblock. > > Is it safe now to run tune2fs on a mounted busy fs? afaik it would > need at least support to quiescence the fs temporarily. Otherwise you > have a race window where changes to the superblock could get lost. Modern versions of e2fsprogs do byte-level writes only to the fields that are being changed by the userspace program; given that these are fields that are not touched by the kernel once the filesystem is mounted, this is safe. - Ted