From: Joseph Fannin Subject: EXT4-fs error: bad header in inode: unexpected eh_depth Date: Fri, 5 Dec 2008 19:28:39 -0500 Message-ID: <20081206002838.GA16688@nineveh.local> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii To: linux-ext4@vger.kernel.org Return-path: Received: from qw-out-2122.google.com ([74.125.92.25]:64465 "EHLO qw-out-2122.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757822AbYLFA3Z (ORCPT ); Fri, 5 Dec 2008 19:29:25 -0500 Received: by qw-out-2122.google.com with SMTP id 3so72797qwe.37 for ; Fri, 05 Dec 2008 16:29:24 -0800 (PST) Content-Disposition: inline Sender: linux-ext4-owner@vger.kernel.org List-ID: Hello, I'm not sure I can reproduce this, but I've seen it twice in two days. I'm hoping someone could tell me a bit about what this means: EXT4-fs error (device dm-1): ext4_ext_search_right: bad header in inode \ #29753858: unexpected eh_depth - magic f30a, entries 2, max 340(0), depth 1(2) ... and then the journal aborts. I ran fsck on the fs after the error both times. After the second one: Log of fsck -C -R -A -a -f Thu Dec 4 18:03:53 2008 home: recovering journal home: Inode 7, i_blocks is 5768, should be 5408. FIXED. home: 105120/37994496 files (10.6% non-contiguous), 74285453/75980800 blocks fsck died with exit status 1 ... nothing about inode #29753858, which is a raw image of a hard disk I was making with `dd` when the errors happened. The first error happened when 17GB had been copied, and the second attempt failed at 27GB -- which don't seem like "interesting" file sizes to me. I was able to make a full 30GB image of the drive today (still inode #29753858), however, and then a copy of that. I am running a 2.6.27.7 kernel with the 20 "for-stable" patches Ted Ts'o sent to stable@kernel.org on Nov 16 applied, which I think is what's in 2.6.27.8 today. Testing .28-rc wouldn't mean much unless I can find a more reliable way to reproduce this. I don't see any other disk- or fs-related errors in my logs other than the fallout from the aborted journal. Could this be caused by a problem with the underlying block device? Is it likely to be disk corruption that's going undetected by fsck or something? Is the extent_header actually *in* inode #29753858 (I'm not sure that question makes even makes sense)? Thanks for any help or info. This filesystem was originally created as ext3, but most of the files on it have been copied or created since it's been converted to ext4. It was created with 256B inodes. dumpe2fs 1.41.3 (12-Oct-2008) Filesystem volume name: home Last mounted on: Filesystem UUID: 4ff8ee4b-b54d-4085-9622-3fa3c6fce4ef Filesystem magic number: 0xEF53 Filesystem revision #: 1 (dynamic) Filesystem features: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize Filesystem flags: signed_directory_hash test_filesystem Default mount options: user_xattr acl Filesystem state: clean Errors behavior: Remount read-only Filesystem OS type: Linux Inode count: 37994496 Block count: 75980800 Reserved block count: 0 Free blocks: 328372 Free inodes: 37915916 First block: 0 Block size: 4096 Fragment size: 4096 Reserved GDT blocks: 45 Blocks per group: 32768 Fragments per group: 32768 Inodes per group: 16384 Inode blocks per group: 1024 Filesystem created: Wed Jun 13 22:44:28 2007 Last mount time: Thu Dec 4 18:07:17 2008 Last write time: Thu Dec 4 18:07:17 2008 Mount count: 1 Maximum mount count: -1 Last checked: Thu Dec 4 18:03:54 2008 Check interval: 0 () Reserved blocks uid: 0 (user root) Reserved blocks gid: 0 (group root) First inode: 11 Inode size: 256 Journal inode: 8 Default directory hash: half_md4 Directory Hash Seed: 1eb54f7f-8b56-4c3b-9120-d0c36f121d0d Journal backup: inode blocks Journal size: 128M -- Joseph Fannin jfannin@gmail.com