From owner-freebsd-current@FreeBSD.ORG Wed Jun 23 15:32:39 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5DD8A16A4CE; Wed, 23 Jun 2004 15:32:39 +0000 (GMT) Received: from spider.deepcore.dk (cpe.atm2-0-53484.0x50a6c9a6.abnxx9.customer.tele.dk [80.166.201.166]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9095943D48; Wed, 23 Jun 2004 15:32:38 +0000 (GMT) (envelope-from sos@DeepCore.dk) Received: from DeepCore.dk (sos.deepcore.dk [194.192.25.130]) by spider.deepcore.dk (8.12.11/8.12.10) with ESMTP id i5NFWVjv080031; Wed, 23 Jun 2004 17:32:36 +0200 (CEST) (envelope-from sos@DeepCore.dk) Message-ID: <40D9A28F.6020602@DeepCore.dk> Date: Wed, 23 Jun 2004 17:32:31 +0200 From: =?ISO-8859-1?Q?S=F8ren_Schmidt?= User-Agent: Mozilla Thunderbird 0.5 (X11/20040329) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Maxim Sobolev References: <5.2.1.1.0.20040621135534.068fef40@server01.chassis00.xsnetworks.net> <40D7E0B7.1000905@portaone.com> <40D8D217.5000808@freebsd.org> <40D95FDC.9060802@portaone.com> <40D99D3B.8010608@DeepCore.dk> <40D99F19.5010008@portaone.com> In-Reply-To: <40D99F19.5010008@portaone.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit X-mail-scanned: by DeepCore Virus & Spam killer v1.4 cc: freebsd-scsi@freebsd.org cc: freebsd-current@freebsd.org cc: Eric Veraart Subject: Re: HostRaid support on FreeBSD X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 23 Jun 2004 15:32:39 -0000 Maxim Sobolev wrote: > Søren Schmidt wrote: > >> Maxim Sobolev wrote: >> >>> Well, my knowelege about internals of software RAID controllers is >>> quite limited, but as long as I know they are no more than BIOS with >>> UI for setting up RAIDs and interface to main BIOS that hides >>> internal RAID structure allowing to boot off the RAID. Also they >>> provide some persistent storage for RAID configuration, I assume >>> either in some controller's NVRAM or on the HDDs. Therefore, the same >>> set of GEOM RAID modules can work with different software RAID >>> controllers (SCSI, ATA, SATA etc) if there will be a standard way to >>> extract this info from the controller's driver, and present it in >>> some common format. That's what I am talking about. >> >> >> >> Well, I dont know about SCSI, but ATA/SATA always stores the metadata >> on disk, the controller has nothing todo with the RAID functionality, >> its merely a way to talk to the disks. >> >> Reading the metadata off the disks and using that info to access a >> RAID array is easy and can be had even without docs in a few hours. >> Writing back the metadata in case of failures etc to retain sanity of >> the array is "somewhat" more difficult. Actually I'm working with this >> currently for new formats for ata-raid and its not trivial to get right. > > > But I assume that on-disk format of this info and its location is > controller-dependent, so that appropriate code to read/decode this info > should probably be in the controller's driver. Not exactly, its BIOS vendor depended, I know of at least a handfull of controller chips that can be had with at least 2 different BIOS software RAID's that uses different metadata. In some cases you can even take a stack of disks thats normally on a HW RAID controller (like the promise sx6000) and use it on a software RAID setup as the metadata is the same and known by ata-raid. At any rate, this isn't as simple as it sounds (as usual) :) -- -Søren