From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 9 Nov 2001 09:44:58 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 9 Nov 2001 09:44:38 -0500 Received: from albatross-ext.wise.edt.ericsson.se ([194.237.142.116]:25497 "EHLO albatross-ext.wise.edt.ericsson.se") by vger.kernel.org with ESMTP id ; Fri, 9 Nov 2001 09:44:30 -0500 Message-ID: <3BEBEBDD.12B8B9FD@eed.ericsson.se> Date: Fri, 09 Nov 2001 15:44:45 +0100 From: "Ronny Lampert (EED)" Organization: Ericsson EuroLab X-Mailer: Mozilla 4.72 [en] (X11; U; Linux 2.2.19 i586) X-Accept-Language: en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: Re: 2.4.14: crashing on heavy swap-load with SmartArray (cont.) In-Reply-To: <3BEBD6E9.3F7F8057@eed.ericsson.se> <20011109143720.T4946@suse.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hello again, thanks a lot for your fast help. Unfortunately this will cause the system to be TOTAL unresponsive and the disks to pause for 10secs regularily not doing anything. No login possible or any interactive work. I saw no process being killed and it's runnig. A ps aux took 5 minutes and still won't finish and kernel-time was 100% on both CPU. The only thing it would bring back to life is SysRq-i (kill all tasks). Are spinlocks the only possibility? This really needs further investigation, it should really be able to handle big loads from matlab-computations usually done by SUNs. Regards, Ronny