From owner-freebsd-stable@FreeBSD.ORG  Sat Feb 13 09:31:34 2010
Return-Path: <owner-freebsd-stable@FreeBSD.ORG>
Delivered-To: freebsd-stable@freebsd.org
Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34])
	by hub.freebsd.org (Postfix) with ESMTP id C2144106568B;
	Sat, 13 Feb 2010 09:31:34 +0000 (UTC)
	(envelope-from danny@cs.huji.ac.il)
Received: from kabab.cs.huji.ac.il (kabab.cs.huji.ac.il [132.65.16.84])
	by mx1.freebsd.org (Postfix) with ESMTP id 7CCAE8FC08;
	Sat, 13 Feb 2010 09:31:34 +0000 (UTC)
Received: from pampa.cs.huji.ac.il ([132.65.80.32])
	by kabab.cs.huji.ac.il with esmtp
	id 1NgELc-00061N-WA; Sat, 13 Feb 2010 11:31:33 +0200
X-Mailer: exmh version 2.7.2 01/07/2005 with nmh-1.2
To: freebsd-hackers@freebsd.org, freebsd-stable@freebsd.org
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Sat, 13 Feb 2010 11:31:32 +0200
From: Daniel Braniss <danny@cs.huji.ac.il>
Message-ID: <E1NgELc-00061N-WA@kabab.cs.huji.ac.il>
Cc: 
Subject: NFS/UDP and vfs.nfs.nfs_ip_paranoia=0 does not help
X-BeenThere: freebsd-stable@freebsd.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Production branch of FreeBSD source code <freebsd-stable.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>, 
	<mailto:freebsd-stable-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-stable>
List-Post: <mailto:freebsd-stable@freebsd.org>
List-Help: <mailto:freebsd-stable-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>,
	<mailto:freebsd-stable-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Sat, 13 Feb 2010 09:31:34 -0000

Hi,
	While trying to find out why our NSF/ZFS servers now hangs about once a
week, I got hold of a similiar box, and got a bit more ambitious, I connected
it via 2 NICs, to complicate things a bit, the server boots via pxeboot (ie, is
datatless). After fiddling with the default gateway, adding -h to rpcbind and
mountd, things seem ok, but UDP is 'problematic', I could do with TCP
except that am-utils does a fsinfo via UDP when doing a /net/<host> and will 
hang
the client.
even with vfs.nfs.nfs_ip_paranoia=0, when the response from the server
arrives with the 'wrong' ip, an ICMP destination unreachable (port unreachable)
is replied.
in short, on the client:
	this works: mount_nfs -o mntudp server-ip-vlanA:/mnt /mnt
	this fails: mount_nfs -o mntudp server-ip-vlanB:/mnt /mnt
since the response is coming from server-ip-vlanA.

Q: why does this work for TCP and fails for UDP
Q: is there a workaround?

danny