From owner-freebsd-questions@FreeBSD.ORG Thu Feb 17 11:42:00 2005 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C56CA16A4CE for ; Thu, 17 Feb 2005 11:42:00 +0000 (GMT) Received: from lorna.circlesquared.com (host217-45-219-85.in-addr.btopenworld.com [217.45.219.85]) by mx1.FreeBSD.org (Postfix) with ESMTP id 92DEC43D41 for ; Thu, 17 Feb 2005 11:41:59 +0000 (GMT) (envelope-from peter@circlesquared.com) Received: from localhost.circlesquared.com (localhost.circlesquared.com [127.0.0.1])j1HBfuxN095762; Thu, 17 Feb 2005 11:41:57 GMT (envelope-from peter@circlesquared.com) From: Peter Risdon To: Simonas Kareiva In-Reply-To: <6035a596050215042329e9c5@mail.gmail.com> References: <6035a596050215042329e9c5@mail.gmail.com> Content-Type: text/plain Date: Thu, 17 Feb 2005 11:41:56 +0000 Message-Id: <1108640516.23699.431.camel@lorna.circlesquared.com> Mime-Version: 1.0 X-Mailer: Evolution 2.0.3 FreeBSD GNOME Team Port Content-Transfer-Encoding: 7bit cc: "freebsd-questions@freebsd.org" Subject: Re: NFS hangs on 5.3-RELEASE-p5 X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Feb 2005 11:42:00 -0000 On Tue, 2005-02-15 at 14:23 +0200, Simonas Kareiva wrote: [...] > > The problem is, that the nfs server hangs after running for a while, > like, 20 minutes. Any file operations (at the nfs mount points and > below) hang on the ftp server too, making it inaccessible. I'm having the same problem here, also with 5.3 but in my case with rsync. Oddly, the hang often happens at exactly the same point in a file hierarchy and I initially suspected a problem with some specific files, then with certain file types (it always seemed to hang when copying pdfs). But it now seems to be more general than that. In fact, even periods of inactivity can cause the nfs mount on the client to time out. > The nfsd > processes on the nfs server do not respond to any signals sent by > `kill`, so only a reboot of the system is a temporary solution. I've found that a forced umount of the nfs mount on the client also eradicates these processes. > We > have ACPI enabled, NMBCLUSTERS set to 65535, SMP, debug.mpsafenet=0 > and also some kern.ipc.* customizations which, as we assume, do not > affect the behavior of nfsd. Everything is served via 3COM 3C940 > single port, 1000baseT adapter (device sk). The whole dmesg is > available at http://webart.lt/~molotov/dmesg I'm having this problem on a stock system, so doubt that your customisations are to blame. I'm on the point of filing a pr - just trying to establish how easy it is to reproduce the problem. Peter.