[ILUG] Stale NFS filehandles

[John P. Looney]

On Thu, May 18, 2000 at 06:47:09PM +0100, Paul Jakma mentioned:
> On Thu, 18 May 2000, John P. Looney wrote:
> >  Er, but the application open()s a file on an NFS filesystem. 
> ok that's between the client and the kernels VFS layer.

 Where else is a problem with a client application going to crop up ;)

> > If the NFS
> > filesystem is trashed, and the kernel knows it 
> which kernel? the server or client? 

 client kernel.

> > (as in, if the file on the
> > NFS file in question is deleted or moved), then it'll return a "state NFS
> > filehandle" error.
> uhmm.. no. AFAIK the server can invalidate the filehandle next time the
> client tries to use it. continious stale file handle messages is [afaik]
> where a client persists in presenting an old file handle. 

 Indeed. And the solution - bounce the client application. Bouncing a
server is overkill.

> but these are nfs internals, i'm way out of my depth. dave airlie might be
> able to speak authoratively about this.. :) all i wanted to do was suggest
> that the guy reboot the client machine, as ime that's where the problem is
> most likely to be.

 I did a lot of playing around with usermode NFS, during my postgrad
(before I got bored of it, and found FIST, the programmable filesystem) -
for what it does, it does well. But it's a shame that Unix filesystems
aren't network-aware - there is no way of telling an application "Sorry,
that volume is gone off the network". It'll just keep retrying, or quit.

Kate

-- 
"The fool must be beaten with a stick, for an intelligent person 
the merest hint is sufficient"                -- Zen Master Greg