From owner-freebsd-fs@FreeBSD.ORG Sat Dec 19 19:09:05 2009 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 92CFB1065676 for ; Sat, 19 Dec 2009 19:09:05 +0000 (UTC) (envelope-from bfriesen@simple.dallas.tx.us) Received: from blade.simplesystems.org (blade.simplesystems.org [65.66.246.74]) by mx1.freebsd.org (Postfix) with ESMTP id 325D68FC13 for ; Sat, 19 Dec 2009 19:09:04 +0000 (UTC) Received: from freddy.simplesystems.org (freddy.simplesystems.org [65.66.246.65]) by blade.simplesystems.org (8.13.8+Sun/8.13.8) with ESMTP id nBJIlm7m028023 for ; Sat, 19 Dec 2009 12:47:48 -0600 (CST) Date: Sat, 19 Dec 2009 12:47:48 -0600 (CST) From: Bob Friesenhahn X-X-Sender: bfriesen@freddy.simplesystems.org To: freebsd-fs@freebsd.org Message-ID: User-Agent: Alpine 2.01 (GSO 1266 2009-07-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; format=flowed; charset=US-ASCII X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.2.2 (blade.simplesystems.org [65.66.246.90]); Sat, 19 Dec 2009 12:47:48 -0600 (CST) Subject: am-utils/NFS mount lockups in 8.0 X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 19 Dec 2009 19:09:05 -0000 After upgrading my FreeBSD system from FreeBSD 7.2 to 8.0, the am-utils automounter is experiencing difficulty with managing the NFS client mounts to my Solaris 10U8 system. There have never been any difficulties before. This is when using the NFSv3/TCP in the default kernel and not the new NFSv4 implementation. If it matters, this is for NFS exports from a ZFS pool, with an exported filesystem per user. The problem I see is that the initial mount is instantaneous and works great. After the mount times out, the re-mount produces several "NFS timeout" messages to the widow where the accessing program is running. Sometimes this remount succeeds, but if it fails, then the program is left locked up forever waiting for the NFS mount. Meanwhile connectivity between the FreeBSD system and the Solaris system is fine, as illustrated by excellent connectivity with SSH and no lost packets via 'ping'. The amd log file (/var/log/amd.log) shows no sign of any difficulties. The only complaint is due to an occasional attempt to mount /home/svn even though there is no 'svn' user in /etc/passwd. Perhaps this is due to some action of portupgrade since during this time all of the ports are being rebuilt. It is indeed quite curious that the last entry I see in amd.log (prior to entries caused by me killing the amd daemon) is an attempt to mount /home/svn: Dec 18 20:06:36 shaggy amd[651]/map: Trying mount of freddy:/home/svn on /.amd_mnt/freddy/home/svn fstype nfs mount_type non-autofs Dec 18 20:14:20 shaggy amd[651]/warn: WARNING: automounter going down on signal 15 Notice that this "Trying mount" attempt did not obtain a response from the OS and perhaps amd is essentially locked up after that point. >From checking the log file, I see that there are 47 attempts to mount /home/svn since the upgrade to 8.0 and no such attempts with FreeBSD 7.2. It is difficult for me to tell if this is some issue with am-utils, or kernel code. It seems that am-utils was updated for 8.0 and there have been changes in the area of NFS as well in order to accommodate the new NFSv4 implementation. Any ideas? Bob -- Bob Friesenhahn bfriesen@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/