Return-Path: Received: from fieldses.org ([174.143.236.118]:44472 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755034Ab1F0WDR (ORCPT ); Mon, 27 Jun 2011 18:03:17 -0400 Date: Mon, 27 Jun 2011 18:03:16 -0400 From: "J. Bruce Fields" To: Bryan Schumaker Cc: "linux-nfs@vger.kernel.org" Subject: Re: [RFC v2] NFSD: Added basic fault injection Message-ID: <20110627220316.GA11688@fieldses.org> References: <4DD6AF42.5050502@netapp.com> <20110623195526.GA12357@fieldses.org> <4E087ABE.5090108@netapp.com> <4E08E1A0.9020101@netapp.com> Content-Type: text/plain; charset=us-ascii In-Reply-To: <4E08E1A0.9020101@netapp.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On Mon, Jun 27, 2011 at 04:01:36PM -0400, Bryan Schumaker wrote: > On 06/27/2011 08:42 AM, Bryan Schumaker wrote: > > On 06/23/2011 03:55 PM, J. Bruce Fields wrote: > >> On Fri, May 20, 2011 at 02:13:22PM -0400, Bryan Schumaker wrote: > >>> I have been looking at adding fault injection to the NFS server in order > >>> to test the client's state manager and recovery threads. Simulating > >>> errors on the server seems like the easiest way to guarentee that > >>> specific errors happen, and happen when we tell them to. > >>> > >>> This patch uses debugfs to add a simple framework for fault injection to > >>> the server. This framework is a config option, and can be enabled > >>> through CONFIG_NFSD_FAULT_INJECTION. Assuming you have debugfs mounted > >>> to /sys/debug, a set of files will be created in /sys/debug/nfsd/. > >>> Writing to any of these files will cause the corresponding action and > >>> write a log entry to dmesg. > >>> > >>> Changes in v2: > >>> - Replaced "forget all state owners" with "forget all open owners" > >>> - Include fs/nfsd/fault_inject.c in the patch > >>> > >>> Suggestions? Comments? > >> > >> If this is still the most recent version of the patch, and if you can > >> confirm that this has been tried and found useful for testing the > >> client, then I'll go ahead and apply for 3.1. > > > > Yes, this is the most recent version of the patch. It is part of my Jenkins setup, but I haven't been running it on a regular basis recently. > > Trond and I just talked about more things to add in for fault injection, so I'll hopefully have a new version sometime in the near future. OK, I'll hold off applying this for now. --b.