From owner-freebsd-hardware@FreeBSD.ORG Sat Dec 18 04:22:34 2004 Return-Path: Delivered-To: freebsd-hardware@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 546FC16A4CE for ; Sat, 18 Dec 2004 04:22:34 +0000 (GMT) Received: from sccrmhc12.comcast.net (sccrmhc12.comcast.net [204.127.202.56]) by mx1.FreeBSD.org (Postfix) with ESMTP id D8D4743D31 for ; Sat, 18 Dec 2004 04:22:33 +0000 (GMT) (envelope-from jshamlet@comcast.net) Received: from testbed.alexandria.homeunix.net (pcp04630981pcs.gambrl01.md.comcast.net[68.49.59.181]) by comcast.net (sccrmhc12) with ESMTP id <20041218042233012007ueqne>; Sat, 18 Dec 2004 04:22:33 +0000 From: jshamlet@comcast.net To: freebsd-hardware@freebsd.org Date: Fri, 17 Dec 2004 23:22:28 -0500 User-Agent: KMail/1.7 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200412172322.29339.jshamlet@comcast.net> Subject: Data corruption with FreeBSD 5.3 and Intel i810/ICH based board X-BeenThere: freebsd-hardware@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: General discussion of FreeBSD hardware List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 18 Dec 2004 04:22:34 -0000 Guys/Gals, This is a follow-on to a question I posted in -questions, as I have done some more research. To recap, I am attempting to install 5.3-REL from CD to a HP Pavilion XE746. I have upgraded the CPU from a celeron to a 850MHz PIII, and bumped the RAM up to 256MB. To simplify things, I have returned to a stock kernel. I originally tried to install using the stock 24x CD-RW drive (master drive on second port) in the unit, to a 2GB ATA drive (master drive on primary port). This did not go terribly well. Several times, when attempting to commit the slice/partition information to the disk, the disk would quit responding, forcing me to restart the installation. After finally getting through the partitioning, the actual install failed several times, claiming that the volume was full. (the /rescue folder filled the entire volume - and was 95MB on the corrupted disk) After many attempts, I finally got a minimal system installed on the disk - only to have it finally go belly-up during the first boot. Many configuration files contained random garbage, and I eventually scrapped the install. Also, using cp to copy files to/from the disk caused corruption. For example, doing a "cp -Rp /rescue /mnt/rescue" resulted in the same deal (volume full at 109%). However, I found that I could use tar to move files successfully. I tested this with the above "rescue" folder, and indeed, I ended up manually creating an install this way from an old SCSI disk that I installed 5.3 on. (note, the install to the SCSI disk on an old Adaptec 2940AU also went flawless, if terribly slowly) Long story short - I installed a Promise TX2 UltraATA controller and attempted the install again. The install went flawlessly (I'm using kmail on the install right now). cp doesn't cause any corruption problems, and the system seems to be running well. I would say this absolves the disk of blane, since it's the same disk as before. Also, and this may be unrelated, but the system is configured to turn ACPI on at all times. (there are no knobs to turn it off). Lastly, I have verified that the image is good (or at least good enough to install a working setup on an old 440LX based PII board) Any ideas what is going on? I have a lot of older PIII i810/815 based systems - and most don't have free PCI slots for an extra ATA controller. Thanks, Seth Henry