Re: [vox-tech] odd nfs problem

[Micah Cowan <micah@cowaMAPSnbox.com>]

On Tue, Dec 05, 2000 at 07:50:41PM -0800, Mark Kim wrote:
> On Tue, 5 Dec 2000, Henry House wrote:
> 
> > On Tue, Dec 05, 2000 at 11:28:45AM -0800, jdnewmil@dcn.davis.ca.us wrote:
> > > google sez... "are you running rpc.statd on the client?"
> > 
> > I wasn't...that helped somewhat. Unfortunately mutt still cannot aquire the
> > lock. It reports: 'Waiting for fcntl lock...'
> 
> Does fcntl lock work under NFS?  I know there is one type of lock that
> doesn't work over NFS...  I thought it was "flock()" call but the man page
> says it's implemented via fcntl.
> 
> Also, could another program possibly have a lock on it and isn't letting
> go?
> 
> -Mark

Sorry - I must not have been watching this thread much.

I've had this exact problem numerous times.  The important thing is
that rpc.lockd must be running on both the mounter and the mountee to use
flock() over nfs.  This runs automatically on my machine from
/etc/rc.d/init.d/nfslock.

-Micah