From: Ted Ts'o Subject: Re: ext4 corruption Date: Mon, 28 Feb 2011 17:55:59 -0500 Message-ID: <20110228225559.GK28617@thunk.org> References: <20110228044307.GQ2924@thunk.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org To: "Bill Huey (hui)" Return-path: Received: from li9-11.members.linode.com ([67.18.176.11]:56217 "EHLO test.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752675Ab1B1W4C (ORCPT ); Mon, 28 Feb 2011 17:56:02 -0500 Content-Disposition: inline In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On Mon, Feb 28, 2011 at 12:18:16PM -0800, Bill Huey (hui) wrote: > What ever is in the latest Ubuntu maverick, 2.6.35 for the kernel. > > The only thing that makes it not looking like it was a storage problem > is this line that comes before the SATA error: > > ----- > Feb 25 04:41:26 finfin kernel: [206290.181230] JBD: Spotted dirty > metadata buffer (dev = sdb1, blocknr = 0). There's a risk of > filesystem corruption in case of system crash. > ----- There are a few places where we update the superblock bypassing the journal layer. (For example, when we set the RO_COMPAT_LARGE_FILE feature flag if it wasn't previously set). Those should be cleaned up, but it's not related to the rest of the scary-looking corruption which you saw. The worst that might happen is specific superblock update might get lost (i.e., the RO_COMPAT_LARGE_FILE feature flag) on a crash before we commit some other superblock change to the journal. - Ted