From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bas Bloemsaat Subject: Re: SATA_SX4: data corruption with simultaneous writes Date: Wed, 29 Dec 2004 09:31:07 +0100 Message-ID: <7c4668e5041229003166ebfc2e@mail.gmail.com> References: <20041228134745.GA13912@solarsystems.de> Reply-To: Bas Bloemsaat Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from rproxy.gmail.com ([64.233.170.194]:65350 "EHLO rproxy.gmail.com") by vger.kernel.org with ESMTP id S261352AbUL2IbH (ORCPT ); Wed, 29 Dec 2004 03:31:07 -0500 Received: by rproxy.gmail.com with SMTP id y7so204240rne for ; Wed, 29 Dec 2004 00:31:07 -0800 (PST) In-Reply-To: <20041228134745.GA13912@solarsystems.de> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: linux-ide@vger.kernel.org FWIW, we've experienced similar problems with software raid systems on this controller. In two machines, with 4 and 3 disks connected to the controllers, WD Raptors, 36.7GB. Tested both ReiserFS (which I unjustly blamed at first) and ext3. Kernels 2.6.7-2.6.9 Funny thing (or not so funny actually) was that after a while the when /dev/md0 (sdb-c-d-e) got stressed bad enough fs corruption got worse and even kernel panics would occur. That's strange, because the system disk was on /dev/sda, which was not even on that controller. But this was only when stressed, so it strikes me as related to the parent call. Anyway, the specs of these machines were pretty similar to Christian's. P4 (2.8 and 3.2) on intel boards (D865GBFL I believe). 1GB mem in each machine. I haven't been able to dive into this, as the machines are at work and, well, working now. Without the disks on those controllers, that is. On Tue, 28 Dec 2004 14:47:45 +0100, Christian Vogel wrote: > I want to report data corruption that is happening here with a > promise sx4 SATA controller during simultaneous writes to four connected > disks, using kernel 2.6.9. > *snip*