Ext3 users - Apr 2005 - short read while checking ext3 journal

My UPS failed and my server took an 'unscheduled outage' a few weeks
ago.
The only casualty appears to be a volume I used for backup.  I usually
maintain data on multiple hard disks, but in this case I errantly had some
data (of marginal value) on this file system.

At this point, the data is not worth enough for me to send the drive out
for data recovery, but it's worth enough to make me want to try to recover
it.

Here's what's happening:
[root at marvin root]# mount /dev/hdb1 /backup
mount: wrong fs type, bad option, bad superblock on /dev/hdb1,
       or too many mounted file systems

[root at marvin root]# fsck /dev/hdb1
fsck 1.32 (09-Nov-2002)
e2fsck 1.32 (09-Nov-2002)
fsck.ext2: Attempt to read block from filesystem resulted in short read
while checking ext3 journal for /backup

I thought removing the journal might help.  I doubt that the files that
interest me were open at the time.

[root at marvin root]# tune2fs -O ^has_journal /dev/hdb1
tune2fs 1.32 (09-Nov-2002)
The needs_recovery flag is set.  Please run e2fsck before clearing
the has_journal flag.


Any ideas?

Thanks!

Tom

[root at marvin download]# ide-smart /dev/hdb1
Id=  3  Status=39  {Prefailure  Online }  Value=219  Threshold= 63  Passed
Id=  4  Status=50  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=  5  Status=51  {Prefailure  Online }  Value=240  Threshold= 63  Passed
Id=  6  Status= 1  {Prefailure  OffLine}  Value=253  Threshold=100  Passed
Id=  7  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=  8  Status=39  {Prefailure  Online }  Value=250  Threshold=187  Passed
Id=  9  Status=50  {Advisory    Online }  Value=208  Threshold=  0  Passed
Id= 10  Status=43  {Prefailure  Online }  Value=253  Threshold=157  Passed
Id= 11  Status=43  {Prefailure  Online }  Value=253  Threshold=223  Passed
Id= 12  Status=50  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=192  Status=50  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=193  Status=50  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=194  Status=50  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=195  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=196  Status= 8  {Advisory    OffLine}  Value=149  Threshold=  0  Passed
Id=197  Status= 8  {Advisory    OffLine}  Value=240  Threshold=  0  Passed
Id=198  Status= 8  {Advisory    OffLine}  Value=145  Threshold=  0  Passed
Id=199  Status= 8  {Advisory    OffLine}  Value=199  Threshold=  0  Passed
Id=200  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=201  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=202  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=203  Status=11  {Prefailure  Online }  Value=253  Threshold=180  Passed
Id=204  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=205  Status=10  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=207  Status=42  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=208  Status=42  {Advisory    Online }  Value=253  Threshold=  0  Passed
Id=209  Status=36  {Advisory    OffLine}  Value=195  Threshold=  0  Passed
Id= 99  Status= 4  {Advisory    OffLine}  Value=253  Threshold=  0  Passed
Id=100  Status= 4  {Advisory    OffLine}  Value=253  Threshold=  0  Passed
Id=101  Status= 4  {Advisory    OffLine}  Value=253  Threshold=  0  Passed
OffLineStatus=2 {Completed}, AutoOffLine=No, OffLineTimeout=5 minutes
OffLineCapability=91 {Immediate Auto SuspendOnCmd}
SmartRevision=16, CheckSum=219, SmartCapability=3 {SaveOnStandBy AutoSave}


[root at marvin download]# uname -a
Linux marvin 2.4.21-15.EL #1 Sun May 16 02:45:06 EDT 2004 i686 athlon i386
GNU/Linux

Can you mount it using an alternate superblock?

On Apr 09, 2005  16:07 -0400, Tom Cooper wrote:
> I thought removing the journal might help.  I doubt that the files that
> interest me were open at the time.
Sounds like the journal has some of the blocks set to zero (i.e. sparse),
which is an error, or you are actually getting read errors in the journal.
> [root at marvin root]# tune2fs -O ^has_journal /dev/hdb1
> tune2fs 1.32 (09-Nov-2002)
> The needs_recovery flag is set.  Please run e2fsck before clearing
> the has_journal flag.
'debugfs -R "feature needs_recovery" /dev/hdb1' is what you
want.  That
should probably be put into the tune2fs man page.

After this you should probably run badblocks before recreating the journal.

Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.

Hi,

On Sat, 2005-04-09 at 21:07, Tom Cooper wrote:
> [root at marvin root]# fsck /dev/hdb1
> fsck 1.32 (09-Nov-2002)
> e2fsck 1.32 (09-Nov-2002)
> fsck.ext2: Attempt to read block from filesystem resulted in short read
> while checking ext3 journal for /backup
Are you getting any kernel IO errors logged during that part of the
fsck?
> I thought removing the journal might help.  I doubt that the files that
> interest me were open at the time.
> 
> [root at marvin root]# tune2fs -O ^has_journal /dev/hdb1
> tune2fs 1.32 (09-Nov-2002)
> The needs_recovery flag is set.  Please run e2fsck before clearing
> the has_journal flag.
> Any ideas?
Yes, you can get around that.  "man tune2fs":

       -f     Force  the  tune2fs  operation  to  complete even in the face of
              errors.  This option is useful  when  removing  the  has_journal
              filesystem feature from a filesystem which has an external jour-
              nal (or is corrupted such that it appears to  have  an  external
              journal), but that external journal is not available.

              WARNING:  Removing  an  external journal from a filesystem which
              was not cleanly unmounted without first replaying  the  external
              journal  can  result  in severe data loss and filesystem corrup-
              tion.

Cheers,
 Stephen