"stale NFS file handle" problem

[Anne Wilson]

On Friday 30 January 2009 16:01:14 David Jansen wrote:
> I have had a lot of nfs instability lately, including the stale nfs file
> handles. The latest update that came out a couple of days ago, seems to
> have fixed this, nfs-utils-1.1.4-7.fc10 . Only, if nfs was already
> messed up on a server, only a reboot seems to help, as far as I have
> been able to figure out.
> So first upgrade to the latest nfs-utils, see if a 'service nfs restart'
> is succesful, and if not, try rebooting. It got things going for me on a
> lot of systems (like a 32 node cluster used for computations, where each
> of the nodes needs to see the disks of the other nodes, so lots of
> cross-mounts, and lots of heavy data access. so if there stil was a bug,
> I guess I would have noticed it there).

The server has had a lot of problems lately - discussed in other threads, 
particularly on the CentOS list.  I'm of the opinion that this message came 
when the problems first started.  I'm guessing that something crashed leaving 
a lock file behind.  After a couple of days of problems the system crashed so 
badly that it could not be recovered.  To my untutored eye, the inode table 
was lost.  I had to reinstall.

Anne
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
URL: <http://listman.redhat.com/archives/fedora-list/attachments/20090130/146ab9fe/attachment-0001.sig>