Help on ext3 file system corruption issue

[Stephen Samuel]

On 6/5/07, Andreas Dilger <adilger at clusterfs.com> wrote:
>
> On Jun 05, 2007  22:01 +0530, Somasundaram, Arun (IE10) wrote:
> > I have a Kernel 2.4.7-10 with ext3 file system in compact flash. The
> > system was up for 3 months and was running with average load conditions.
>
> Unless you have a support contract with some vendor, nobody will look at
> bugs from such an old kernel.  There are a hundred old bugs that might
> have been fixed already.


There are extensions to DD that will, on an error allow you to skip over the
block(s) in error while zeroing (instead of just ignoring) the blocks on the
output... (I think that Knoppix has such a version of DD, if that'll help
you)  This means that everything that can be read will be where (relative to
the start of your recovery partition or file ) ext3fs is  expecting to  find
it.   Once you find the recovery DD, use it to copy your filesystem to a
hard drive or whatever, You can then recover your data and then
--- presuming andreas is right -- you'll have to replace your flash, and
then put a filesystem on it that's more conducive to how flash works.


> hda: read_intr: status=0x51 { DriveReady SeekComplete Error }
> >
> > hda: read_intr: error=0x40 { UncorrectableError }, LBAsect=667854,
> > sector=163854
> >
> > end_request: I/O error, dev 03:02 (hda), sector 163854
> >
> > EXT3-fs error (device ide0(3,2)): ext3_get_inode_loc: unab
> >
> > le to read inode block - inode=20089, block=81926
>
> This is likely a hardware error.  Probably due to the fact that ext3
> is not a good filesystem to use on CF because the journal is always
> overwriting the same part of the CF device.  Try something like JFFS2
> instead.
>
> Cheers, Andreas
>
>


-- 
Stephen Samuel http://www.bcgreen.com
778-861-7641
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/ext3-users/attachments/20070605/0a04db1b/attachment.htm>