How do I debug NFS write error on host ABC: Stale NFS file handle?

How do I debug NFS write error on host ABC: Stale NFS file handle?

Post by John Eismeie » Wed, 18 Jul 2001 04:55:54



Hi All,

How do I debug NFS write error on host abc: Stale NFS file handle?  I
have tried the fhfind but the script has not found the filesystem.

Thank you.

--

 
 
 

How do I debug NFS write error on host ABC: Stale NFS file handle?

Post by Lon Stowe » Wed, 18 Jul 2001 11:52:04




Quote:

>Hi All,

>How do I debug NFS write error on host abc: Stale NFS file handle?  I
>have tried the fhfind but the script has not found the filesystem.

   A stale file handle means you tried to use a file handle that
   no longer exists on the server.   This makes things a bit tough
   for fhfind to decode.  

 
 
 

How do I debug NFS write error on host ABC: Stale NFS file handle?

Post by Mathew Kirsc » Thu, 19 Jul 2001 04:52:07



> How do I debug NFS write error on host abc: Stale NFS file handle?  I
> have tried the fhfind but the script has not found the filesystem.

I usually know why I get "Stale NFS file handle" messages, and debugging is
not required. It's almost always after an NFS server gets rebooted
(intentionally or otherwise). I believe it has to do with the fact that some
process was either using a directory within the NFS mount as a present working
directory (as $PWD), and/or some process is using a file within the NFS mount.
The fix is to umount and remount the directory if it's a static NFS mount, or
simply umount it if it's automounted.

What are the circumstances leading up to this problem?

 
 
 

How do I debug NFS write error on host ABC: Stale NFS file handle?

Post by Marvi » Thu, 19 Jul 2001 14:27:15




> > How do I debug NFS write error on host abc: Stale NFS file handle?  I
> > have tried the fhfind but the script has not found the filesystem.

> I usually know why I get "Stale NFS file handle" messages, and debugging is
> not required. It's almost always after an NFS server gets rebooted
> (intentionally or otherwise). I believe it has to do with the fact that some
> process was either using a directory within the NFS mount as a present working
> directory (as $PWD), and/or some process is using a file within the NFS mount.
> The fix is to umount and remount the directory if it's a static NFS mount, or
> simply umount it if it's automounted.

> What are the circumstances leading up to this problem?

Did you check nfsstat ?
 
 
 

How do I debug NFS write error on host ABC: Stale NFS file handle?

Post by Steve Boyl » Thu, 19 Jul 2001 22:11:28


The message indicates that it can no longer find a file handle it was
working
with.

 This AFAIK usually occurs when either an open file is somehow removed or
moved (this also generates those .nfsXXXX files you see lying about your
filesytems) or the exported filesystem has changed on the server.

For instance, if you take down your nfs server and move /export to another
device it's file handle will change. When you reboot the system, any client
that was accessing the old filehandle will now complain about stale nfs
filesystems.

The best bet to avoid this happening is to unmount the filesystems on the
clients before changing any filesystems on the server.

Cheers
  Steve



>> How do I debug NFS write error on host abc: Stale NFS file handle?  I
>> have tried the fhfind but the script has not found the filesystem.

>I usually know why I get "Stale NFS file handle" messages, and debugging is
>not required. It's almost always after an NFS server gets rebooted
>(intentionally or otherwise). I believe it has to do with the fact that some
>process was either using a directory within the NFS mount as a present
working
>directory (as $PWD), and/or some process is using a file within the NFS
mount.
>The fix is to umount and remount the directory if it's a static NFS mount, or
>simply umount it if it's automounted.

>What are the circumstances leading up to this problem?

*********************

*********************
 
 
 

1. Diskless NFS error - "Stale NFS File Handle"

I am trying to configure a terminal to do a diskless/root-NFS mount of my
server. Both systems are using RedHat 6.1. The problem I have is :

I can get the kernel to load up to the start of starting rc3.d/S25netfs. At
this point the system tries to mount an NFS partition from my fstab and the
system hangs. I get two errors "Stale NFS file handle"

I can get around the problem by removing the script (just renamed it to
s25netfs.) I can the log in as root and execute programs. As soon as I try to
mount another nfs partition I loose my root-NFS parition and all I get with
ANY command from then on is "Stale NFS file handle"

Any ideas or suggestions?

Chris

PS - If any details are missing, please let me know. It is very important that
I get this working.

2. PCL to PDF

3. 2.6.0 NFS server giving 'stale NFS handle' errors

4. automatic tasks

5. nfs-mount: stale NFS file handle

6. metathemes

7. Stale NFS file handle for NFS mounted directory

8. See this for help (in Danish)

9. NFS: Stale File Handle error

10. cat produces stale NFS file handle error!?

11. Stale NFS File handle error on /mnt/cdrom

12. Need help matching stale NFS file handle to a file, Solaris 2.5

13. NFS Stale file handle....