From owner-freebsd-amd64@FreeBSD.ORG Fri Feb 24 18:42:04 2006 Return-Path: X-Original-To: freebsd-amd64@freebsd.org Delivered-To: freebsd-amd64@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8BC2916A420 for ; Fri, 24 Feb 2006 18:42:04 +0000 (GMT) (envelope-from scottl@samsco.org) Received: from pooker.samsco.org (pooker.samsco.org [168.103.85.57]) by mx1.FreeBSD.org (Postfix) with ESMTP id 0F23A43D45 for ; Fri, 24 Feb 2006 18:42:03 +0000 (GMT) (envelope-from scottl@samsco.org) Received: from [10.10.3.185] ([69.15.205.254]) (authenticated bits=0) by pooker.samsco.org (8.13.4/8.13.4) with ESMTP id k1OIfm5p011681; Fri, 24 Feb 2006 11:41:51 -0700 (MST) (envelope-from scottl@samsco.org) Message-ID: <43FF5366.5030002@samsco.org> Date: Fri, 24 Feb 2006 11:41:42 -0700 From: Scott Long User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060206 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "V. T. Mueller" References: <3.0.1.32.20060222094147.00a928a0@pop.redshift.com> <20060224021327.H1005@ganymede.hub.org> <20060224064125.GA1782@xor.obsecurity.org> <43FEC2D9.7090407@samsco.org> <43FEDBC4.3010103@datafarm.de> In-Reply-To: <43FEDBC4.3010103@datafarm.de> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Status: No, score=0.0 required=3.8 tests=none autolearn=failed version=3.1.0 X-Spam-Checker-Version: SpamAssassin 3.1.0 (2005-09-13) on pooker.samsco.org Cc: eniazi@ossmail.com, freebsd-amd64@freebsd.org, Kris Kennaway Subject: Re: anyone successfully using an icp-vortex / iir controller? X-BeenThere: freebsd-amd64@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting FreeBSD to the AMD64 platform List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 24 Feb 2006 18:42:04 -0000 V. T. Mueller wrote: > > Hello, > > Scott Long schrieb: > >> Kris Kennaway wrote: >> >>> On Fri, Feb 24, 2006 at 02:13:56AM -0400, Marc G. Fournier wrote: >>> >>>> I run an IIR controller on one of our servers just fine ... 4GB of >>>> RAM, FreeBSD 4.x ... haven't tried 6.x on it yet ... >>> >>> It's known to be broken on any system with >4GB of RAM. >> >> I'll strengthen that statement and say that it is completely, 100% >> unusable with 4GB of RAM or more. You are guaranteed to get data >> corruption on the second I/O that you send. Fixing this requires >> re-writing much of the driver, unfortunately. > > > If this is true - and I think so, I wasted a lot of time on it already - > so why isn't there a little piece of documentation speaking out a > warning. We've been happily using a large number of GDT controllers > under FreeBSD for years after we have had a couple of bad experiences > with other brands. > > One of those, a mylex 160lp was just tested to be unfunctional with > 6.1b2 (8GB RAM), too. > > Are there any *good* SCSI RAID controllers out there that are known to > be working in server setups (amd64, SMP, >4GB RAM)? > > When thinking it over, I guess I have to be grateful that over many > years I never experienced s/th that made me question the "The Power To > Serve" slogan. What I really feel unhappy about right now is that there > was no information available so I could take appropriate measures in > order to avoid the current situation (at least none that I could find). > > Well, what's the plan regarding the iir driver now? I would suspect that > letting folks rendering their servers unfunctional by upgrading memory > is not really a good strategy. > > The manpage lists two authors/maintainers, I emailed both early this > week. Achim Leubner will be back from vacation the 27th. So maybe he can > contribute s/th constructive. > > Kind regards, > vt Well, there are several ways that I can answer these questions. I could say that the iir driver was developed and supported by the vendor, and that you should contact the vendor for support questions like this. I could say that the author of this driver didn't understand how to use FreeBSD's driver API's, and that to a casual reviewer, the driver looks like it should work, and that only with a deeper look do the problems become apparent. I could say that while I personally support the vast majority of the RAID drivers in FreeBSD, I only have the time and resources to actively maintain and fix the most popular ones, as taking care of a single driver for a single architecture usually involves an engineering and testing staff of at least a dozen people at most companies, and I try to do the equivalent job for 8 different drivers, all in my 'free time'. So, I understand your frustration over this. There are a number of other RAID solutions that work quite well in FreeBSD. Adaptec AAC, LSI MegaRAID, IBM/Adaptec ServeRAID, Compaq/HP CISS all are very reliable choices for SCSI RAID. I know that that doesn't help you with your ICP/Vortex choice, and I do sympathize with your problem. I've done my best to modernize as many drivers as I can to work with >4GB, but I've had to prioritize my efforts based on the popularity and the future returns of a given architecture. ICP/Vortex basically died when Adaptec bought them, so there isn't a lot of motivation to put energy into the driver since there won't ever be any new cards from them. However, if someone is interested in fixing this driver, I can help guide them. Scott