[lug] Stale NFS File Handle

John Hernandez John.Hernandez at noaa.gov
Thu Aug 4 13:44:49 MDT 2005


If you have some extra space, consider creating an ext3 filesystem on an
LVM2 volume, then export that and see what happens.  I expect you might
have better luck with ext3 because it's so common.

Dan Ferris wrote:
> Hmmmm.
> 
> Figures this would be my luck.
> 
> Ok, I took the automounter out of the picture on our Sun E450 (Solaris
> 8) and one of the Linux workstations.  The Sun has one error, the Linux
> box seems to be chugging along.
> 
> Hmmmm, this is an interesting problem.
> 
> I'd try a later version of the kernel, but 2.6.12 for FC 4 has a SCSI
> driver that doesn't like one of my RAID arrays.
> 
> Oh yes, the filesystem on the LVM is JFS.  Why JFS?  Because it was
> easiest to resize and it seems to work.
> 
> Dan Ferris
> 
> Hugh Brown wrote:
> 
>> You don't seem to be alone:
>>
>> http://www.redhat.com/archives/fedora-devel-list/2004-November/msg01115.html
>>
>>
>> I have vague memories of reading that LVM2 and NFS don't get along.  Now
>> that I think about it, it was probably the extended acl's and nfs that
>> didn't get along.
>>
>> Can you take automounter out of the picture and see if you still get the
>> errors?
>>
>> Hugh
>>
>> On Thu, 4 Aug 2005, Dan Ferris wrote:
>>
>>  
>>
>>> Already been done.
>>>
>>> It's still happening.
>>>
>>> At the moment, I'm trying to figure out if the automounter is causing
>>> the problem.
>>>
>>> Dan
>>>
>>> Hugh Brown wrote:
>>>
>>>   
>>>
>>>> Worst case, schedule a down time, make sure the clients don't have
>>>> anything mounted, restart the nfs daemon (and friends).
>>>>
>>>> Hugh
>>>>
>>>> On Thu, 4 Aug 2005, Dan Ferris wrote:
>>>>
>>>>
>>>>
>>>>     
>>>>
>>>>> Here's a good one for all the other Linux experts....
>>>>>
>>>>> I just moved a bunch of data on my NFS server over to LVM2.  So we
>>>>> have
>>>>> about 2 TB of data that lives in LVM and it's being served out via
>>>>> NFS.
>>>>>
>>>>> So, the problem now is that some users are getting a "Stale NFS file
>>>>> handle error" when using their data.  Their directories are mounted
>>>>> via
>>>>> the automounter.  The problem seems to show up after a few minutes of
>>>>> the directory being mounted via the automounter.
>>>>>
>>>>> This problem seems random and intermittent (the best kind).
>>>>>
>>>>> The NFS server is Fedora Core 4, although I'm running on an older
>>>>> version of the kernel (2.6.9).
>>>>>
>>>>> If anyone knows how to make that error message go away, I'd appreciate
>>>>> knowing.
>>>>>
>>>>> Dan Ferris
>>>>>       
>>
>> _______________________________________________
>> Web Page:  http://lug.boulder.co.us
>> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
>> Join us on IRC: lug.boulder.co.us port=6667 channel=#colug
>>
>>
>>  
>>
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: lug.boulder.co.us port=6667 channel=#colug
> 

-- 

 |  John Hernandez - NOAA Boulder NOC - 303-497-6392
 |  Mailstop R/OM62. 325 Broadway, Boulder, CO 80305
 |  PGP Public Key ID: 586A7E23



More information about the LUG mailing list