Re: SunOS 5.8 NFS server mounted by linux 2.4.20 client Locking problem

From: Robert Watson <rwatson_at_freebsd.org>
Date: Thu, 26 Feb 2004 10:13:58 -0500 (EST)
On Wed, 25 Feb 2004, Rob Heise (rheise) wrote:

> Any one have a good idea whats going on here:
>  
> open("/usr/SD/tibco/jms/3.1.2/bin/datastore/meta.db",
> O_RDWR|O_CREAT|O_LARGEFILE, 0666) = 7
> fcntl64(7, F_SETLK64, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0},
> 0xbfffe550) = -1 ENOLCK (No locks available)
> close(7)                                = 0
>  
> rpc.lockd and statd are running.  

Well, Solaris generally returns ENOLCK from fcntl() when you exceed a
resource limit on the number of locks permitted on a single file or
region.  However, it looks like you truss is on the client, so it would be
interesting to see what the NFS server returns to the client on the wire.
This isn't really the pervue of the FreeBSD mailing lists, but I've found
a combination of tcpdump and ethereal to be an ideal NFS debugging tool:
ethereal can provide detailed decoding of NFS RPCs that are invaluable in
figuring out what's actually going on, despite what the client may say is
happening.  Remember that the NFS client will often interpose its own
limits, translations of error numbers, etc, in ways that can be
misleading...

Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
robert_at_fledge.watson.org      Senior Research Scientist, McAfee Research
Received on Thu Feb 26 2004 - 06:15:00 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:44 UTC