From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263834AbTETPym (ORCPT ); Tue, 20 May 2003 11:54:42 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263833AbTETPym (ORCPT ); Tue, 20 May 2003 11:54:42 -0400 Received: from [208.186.192.194] ([208.186.192.194]:7896 "EHLO mail.osdl.org") by vger.kernel.org with ESMTP id S263825AbTETPyj (ORCPT ); Tue, 20 May 2003 11:54:39 -0400 Message-Id: <200305201607.h4KG7XL21672@mail.osdl.org> X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: akpm@digeo.com, linux-kernel@vger.kernel.org Subject: wacky hangs on 2.5.69-mm7, -mm3 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 20 May 2003 09:07:33 -0700 From: Cliff White Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Andrew, we are running a few tests on 2.5.69, -mm7, and -mm3 and having some wacky hangs. Booted both kernels with default IO scheduler. The Reaim test running against -mm7 (AIM rewrite) hangs, with a large number of 'sync' commands that never complete and are unkillable. (Various tests call 'sync') System is a 4-CPU PIII , two of the CPU's are stuck at 100% iowait. Filesystem is ext2. Same kernel run with elevator=deadline ran okay. -mm6 ran okay, but was not especially performant. (separate mail) The -mm3 kernel is running DBT2 on a 8-CPU machine, with raw devices. Ten minutes into the run, all IO ceases, iostat shows devices 133% busy, workload stops. 'ps -aef' never completes. Previous run with elevator=noop was fine. We've left the 8-way as is if you want a look. cliffw maryedie OSDL