From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: linux-nfs-owner@vger.kernel.org Received: from plane.gmane.org ([80.91.229.3]:37216 "EHLO plane.gmane.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751598Ab3AFKtE (ORCPT ); Sun, 6 Jan 2013 05:49:04 -0500 Received: from list by plane.gmane.org with local (Exim 4.69) (envelope-from ) id 1Trnn2-0004J7-KS for linux-nfs@vger.kernel.org; Sun, 06 Jan 2013 11:49:16 +0100 Received: from p4FE15490.dip.t-dialin.net ([79.225.84.144]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 06 Jan 2013 11:49:16 +0100 Received: from mario.bachmann by p4FE15490.dip.t-dialin.net with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 06 Jan 2013 11:49:16 +0100 To: linux-nfs@vger.kernel.org From: Mario Bachmann Subject: Re: [PATCH 0/2] Fix regression on NFS in mainline kernel Date: Sun, 6 Jan 2013 10:48:48 +0000 (UTC) Message-ID: References: <20120418105147.2a6f1c55@corrin.poochiereds.net> <1334770614-10653-1-git-send-email-Trond.Myklebust@netapp.com> <20120419093025.GA2916@zeus> <20120419134847.GC2916@zeus> <1334856549.21419.48.camel@lade.trondhjem.org> <20120419204638.GB22948@zeus> <4FA345DA4F4AE44899BD2B03EEEC2FA91198A4B4@SACEXCMBX04-PRD.hq.netapp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-nfs-owner@vger.kernel.org List-ID: Myklebust, Trond writes: > > I use kernel 3.7.1 from kernel.org which has the patches included, > > which should fix the problems (I checked the source code). But: > > I get "NFS: nfs4_reclaim_open_state: Lock reclaim failed!" on the > > client, which mounts home over nfs4. > > I get senseless network traffic (50MB/s on my gigabit network). > > Finally I get Kernel Ops on the Server sometimes. > > Lock reclaiming is not guaranteed to succeed; it all depends on why the > lock was lost in the first place. Did the server restart, or did a > network partition occur? Was some other process able to take that lock > (or even delete the file, doh!), before your client could reclaim it? > > That said, if the server is Oopsing, then you have a broken server. I'd > suggest posting that Oops to this list first. > Thanks for the hint, the server was really broken! The power supply of the server had some broken capacitors. With a new power supply there are no oopses any more! So my reported problems with locking are worthless, because it happened with the broken server.