From mboxrd@z Thu Jan 1 00:00:00 1970 From: Randy Dunlap Subject: Re: data=journal busted Date: Fri, 16 Feb 2007 14:50:22 -0800 Message-ID: <20070216145022.12e86c7d.randy.dunlap@oracle.com> References: <20070215204445.411d2760.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: "linux-ext4@vger.kernel.org" To: Andrew Morton Return-path: Received: from agminet01.oracle.com ([141.146.126.228]:11797 "EHLO agminet01.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751626AbXBPWyW (ORCPT ); Fri, 16 Feb 2007 17:54:22 -0500 In-Reply-To: <20070215204445.411d2760.akpm@linux-foundation.org> Sender: linux-ext4-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Thu, 15 Feb 2007 20:44:45 -0800 Andrew Morton wrote: > > I have a report from a google person who just did some basic > power-it-off-during-a-write testing on 2.6.20's ext3. ordered-data is OK, > but data=journal came back with crap in the file data. > > Is anyone doing any formal recovery stress-testing? > > I suspect we should resurrect and formalise my old > make-the-disk-stop-accepting-writes-when-a-timer-goes-off thing. It was > very useful for stress-testing recovery. Where is your make-troubles patch, please sir? --- ~Randy *** Remember to use Documentation/SubmitChecklist when testing your code ***