[Jfs-discussion] JFS filesystem corruption and jfs_fsck

[Dave Kleikamp]

On Wed, 2009-04-01 at 16:48 -0400, Michael Peek wrote:
> Hi gurus,
> 
> I have a server with a JFS filesystem on it that's gotten corrupted.  
> Maybe I just don't know how to use jfs_fsck, but it doesn't seem to be 
> doing the trick.  It prints out lots of filesystem errors but apparently 
> doesn't repair them.  Attached is a log of the output from jfs_fsck.  
> The exit code from jfs_fsck is 134.

You seem to be using jfs_fsck correctly, but it's finding something it
can't handle.  I don't know why it was written this way, but jfs_fsck
likes to find all the errors first and then fix them all at the end.
Unfortunately, something is preventing it from finishing.

rc=-268 in phase 0 indicates that the journal superblock is
unrecognized.  This could mean that a large part of the volume could be
corrupted.  Could this be caused by a RAID problem, such as a disk going
offline?

I think exit code 134 means that the program was killed by a SIGABRT
signal.  I'm not sure where that's coming from.  Is there anything at
the end of the output that gives a clue?  Is there anything unusual in
the syslog?

> Does anyone have an idea how I can fix this?

If we can't figure out why jfs_fsck is failing, or can't find a way
around it, you could try mounting the volume read-only to recover as
much as you can.

> I'm using the version of jfs that is shipped with ubuntu 8.10 server 
> amd64: 1.1.12-1
> 
> Thanks in advance,
> 
> Michael

-- 
David Kleikamp
IBM Linux Technology Center