Re: [2.6.19] NFS: server error: fileid changed

[Martin Knoblauch <knobi@knobisoft.de>]

--- Trond Myklebust <trond.myklebust@fys.uio.no> wrote:

> On Mon, 2006-12-11 at 08:09 -0800, Martin Knoblauch wrote:
> > Hi, [please CC me, as I am not subscribed]
> > 
> >  after updating a RHEL4 box (EM64T based) to a plain 2.6.19 kernel,
> we
> > are seeing repeated occurences of the following messages (about
> every
> > 45-50 minutes).
> > 
> >  It is always the same server (a NetApp filer, mounted via the
> > user-space automounter "amd") and the expected/got numbers seem to
> > repeat.
> 
> Are you seeing it _without_ amd? The usual reason for the errors you
> see are bogus replay cache replies. For that reason, the kernel is
> usually very careful when initialising its value for the
> XID: we set part of it using the clock value, and part of it
> using a random number generator.
> I'm not so sure that other services are as careful.
>

 So far, we are only seeing it on amd-mounted filesystems, not on
static NFS mounts. Unfortunatelly, it is difficult to avoid "amd" in
our environment.
 
> >  Is there a  way to find out which files are involved? Nothing
> seems to
> > be obviously breaking, but I do not like to get my logfiles filled
> up. 
> 
> The fileid is the same as the inode number. Just convert those
> hexadecimal values into ordinary numbers, then search for them using
> 'ls
> -i'.
> 

 thanks. will check that out.

Cheers
Martin

------------------------------------------------------
Martin Knoblauch
email: k n o b i AT knobisoft DOT de
www:   http://www.knobisoft.de