From mboxrd@z Thu Jan 1 00:00:00 1970 From: bfowler@ewitness.co.uk (ewitness - Ben Fowler) Subject: RAID 5 disk format for Promise SuperTrak Sx6000 (Linux) Date: Sun, 26 May 2002 03:35:09 +0100 Sender: linux-raid-owner@vger.kernel.org Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Return-path: To: linux-raid@vger.kernel.org Cc: ewitness - David Pearce List-Id: linux-raid.ids Just as a standard disk has a standard format with MBR, partition table, superblock, root directory et cetera; I assume that the disks making up a RAID 5 array have a defined format covering the 'superblock', striping, parity and so forth. Is this standard for all RAID compliant/certified systems? Is it defined by each vendor? Is the format/system used by Promise published anywhere? Owing to the the sequential ocurrence of a number of circumstances each of which was distinctly unlikely. I have the job of trying to read 120 GB of data from a 240 GB RAID 5 Array attached to a Supertrak Sx6000 controller card. The software that handles the raid is on the card's BIOS and so far as I know is proprietary. I know that I need help because of the five disks making up the array no fewer than three are recognised as 'free', that is, not part of the array. I suspect that the data on these disks is fine, and that the problem was a transient failure on one of the disks, followed by a catastrophe. If I knew how, I would 'force' the controller to use and resync those disks. I suspect my job would be measurably easier and greatly eased if I could use the standard Linux tools such as mkraid. See which contains the highly optimistic suggestion: The problem was how to read using the Linux RAID 5 software, the data written by Mylex controller. I wrote a few test programs, and found that the Mylex AcceleRaid controller is using what Linux software RAID calls right-asymmetric parity algorithm (you will find the part of /etc/raidtab file I use, at the end of this mail), so I could read the data using Linux software RAID (mkraid --force --dangerous-no-resync /dev/md2). Is there something similar for the Sx6000? How can I find out? Ben.