Re: Severe data corruption with ext4

[Theodore Tso]

On Mon, Mar 23, 2009 at 03:20:03PM +0100, Richard Höchenberger wrote:
> Hi Ted, I just compiled and installed 2.6.29-rc8 vanilla. Before
> running it, I booted from a live CD and fsck'd all my file systems to
> get them into a consistent, clean state.
> 
> Now, after booting 2.6.29-rc8 and writing stuff to dm-14 (/usr), I again get:
> 
> ----------
> Mar 23 15:01:02 bakunin kernel: __find_get_block_slow() failed.
> block=135274289954816, b_blocknr=0
> Mar 23 15:01:02 bakunin kernel: b_state=0x00310021, b_size=4096
> Mar 23 15:01:02 bakunin kernel: device blocksize: 4096
> Mar 23 15:01:02 bakunin kernel: __find_get_block_slow() failed.
> block=135274289954816, b_blocknr=0
> Mar 23 15:01:02 bakunin kernel: b_state=0x00310021, b_size=4096
> Mar 23 15:01:02 bakunin kernel: device blocksize: 4096
> Mar 23 15:01:02 bakunin kernel: grow_buffers: requested out-of-range
> block 135274289954816 for device dm-14
> Mar 23 15:01:02 bakunin kernel: EXT4-fs error (device dm-14):
> ext4_xattr_delete_inode: inode 4501: block 135274289954816 read error
> ----------

Hmm, so the block number is: 7B0800000000 that indicats that
i_file_acl is 0, but the bits in i_file_acl_high are set.  E2fsck
doesn't currently detect and fix this case --- which is a bug in
e2fsck that I'll fix --- but it begs the question how the
i_file_acl_high could have gotten set in the first place.

Was this a freshly created ext4 filesystem, or one that was converted
from ext3?

					- Ted
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/