From: Theodore Tso Subject: Re: Segmentation Faults with both 062608 snapshots Date: Mon, 30 Jun 2008 22:32:53 -0400 Message-ID: <20080701023252.GA28143@mit.edu> References: <3.0.6.32.20080625135340.02423ed8@pop.west.cox.net> <3.0.6.32.20080625135340.02423ed8@pop.west.cox.net> <3.0.6.32.20080626221227.0242af78@pop.west.cox.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Eric Sandeen , "linux-ext4@vger.kernel.org" To: Gary Hawco Return-path: Received: from www.church-of-our-saviour.org ([69.25.196.31]:40637 "EHLO thunker.thunk.org" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752966AbYGACdF (ORCPT ); Mon, 30 Jun 2008 22:33:05 -0400 Content-Disposition: inline In-Reply-To: <3.0.6.32.20080626221227.0242af78@pop.west.cox.net> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Jun 26, 2008 at 10:12:27PM +0000, Gary Hawco wrote: > More on my segmentation problems. Just looking at the end of each kernel > message, they always end with: > EIP [XXXXX] jbd2_journal_data_metadata-XXX Hmm... any chance you can take a picture of the crash message with a digital camera and post the jpg? (I'm assuming you haven't been able to capture the OOPS stack trace in /var/log/kern.log or /var/log/messages or some such.) > This occurs when I try to copy data from my main partition to a backup > partition or make a tarball. I am currently using the ext4 patch queue comit id #555132eb from 2008-06-30 13:04:35 -0400, and I can't reproduce it. I just tried backing up by isync mail directory to a tar.gz file and then restored it, with no problems. Both the source and backup partition are ext4 filesystems with flex_bg, meta_bg, and uninit_bg, located on separate LVM logical volumes on my laptop.[1] [1] http://thunk.org/tytso/blog/2008/06/30/ext4-is-now-the-primary-filesystem-on-my-laptop/ The partition is mounted using mount options: noatime,errors=remount-ro,barrier=1,data=ordered according to /proc/mounts. The big difference between your mount options and mine is that I dont have journal_async_commit as a mount option. Which would be largely moot for me since LVM doesn't support barrier operations. :-( :-( :-( You seem to be able to reproduce the problem at will. Could you try removing te journal_sync_commit option and see if it makes the problem go away? That would be very interesting if it were the case... - Ted