I'm having this problem with NFS that is just giving me a major headache. I have a share /mnt/user/Television which has a bunch of directories. I have NFS, AFP and SMB enabled. I'm using NFS on my linux box and AFP on my mac.

 

Most of the directories work just fine but a few of them give me a stale NFS mount. I've tried unmount/mount and rebooting both machines several times. I've even tried to disable AFP and SMB on the share I'm trying to get NFS to work on.

 

s: cannot access ./The Borgias/.: Stale NFS file handle

ls: reading directory ./The Borgias: Stale NFS file handle

total 12618415

      ? d??? ? ?        ?              ?                ? .

      2 drwxrwx--- 1      99 users      2064 2011-08-14 06:19 ..

      0 drwxrwsr-x 1 99 users        72 2011-07-06 15:29 .AppleDouble

2295488 -rw-rw-r-- 1 99 users 2348285932 2011-03-31 21:06 The.Borgias.s01e01.The.Poisoned.Chalice.mkv

 

...

 

I'm all out of ideas. I'm running beta10.

 

I think somehow the directory got screwed up. I was able to create a new directory on my server, move the folder contents to the new directory and delete the old directory and rename it back to the original name.

 

I'm using this script:

 

#!/usr/bin/perl
use strict;

opendir DIR, "." or die $!;
my @dir = grep {!/^\./ && -d $_ } readdir DIR;
foreach my $i (sort @dir) {
        print qq|echo "$i"\n|;
        print qq|mkdir "2$i"\n|;
        print qq|mv "$i/"* "2$i"\n|;
        print qq|rmdir "$i"\n|;
        print qq|mv "2$i" "$i"\n|;
}
closedir DIR;

 

Everything seems to be working now but I still don't know how it got messed up in the first place.

I had this happen on a ubuntu box and just did:

 

/usr/sbin/exportfs -f

 

on the server to flush out the export table, then another request from the client worked fine. I guess the same might work for unRAID? Perhaps a linux guru could elaborate?

I experienced this problem a few days ago.  In the end I just rebooted and it went away.

My experience with NFS on UnRaid hasn't been great either. I decided to not waste any time and instead just apt-got smbfs on the linux clients and mounted stuff that way

 

Add the following line to /etc/fstab:

 

//tower/data /data   smbfs   noatime,nodiratime,username=anything,password=,uid=1000,gid=1000   0       0

 

where //tower/data obviously is the sharename on UnRaid and /data is the mountpoint. The uid and gid options are required to make the mount userwritable (1000 is the uid for the default user on ubuntu and debian, if you use a different distro, use id to verify)

I experienced this problem a few days ago.  In the end I just rebooted and it went away.

 

I have found reboots are about the only thing that work. I've tried just about everything else. Unfortunately, the problem keeps coming back.

I've been having issues with stale nfs handle as well, my openelec xbmc machine mounts my media share as I was having issues with certain scripts on "non-local" files.  But I'm getting the stale NFS stuff pretty often on that machine when I try to access the directory.  Rebooting the client fixes it temporarily but it comes back fast, I've not tried rebooting the unraid server yet.  I also have two unraid servers running right now, each one mounts the others shares to facilitate transferring files using MC and have one machine running sabnzbd and the final files being moved to the other machine.. That one also has issues with the stale handles.  I've found that unmounting and remounting also fixes it temporary but I'd love to know what's causing it and how to stop it from happening.

There is some type of problem with NFS Stale issues on client side.   I have used VMWare and XenServer as NFS clients for the past couple of years and NFS has been flawless.   Now with 5.0 it is too unstable.

 

Anyone have any suggestions?

 

 

Try removing all traces of the NFS mount from both the clinet and the server and reboot each. Then reconfigure NFS.