From owner-freebsd-isp  Wed Mar  3  1:41:45 1999
Delivered-To: freebsd-isp@freebsd.org
Received: from verdi.nethelp.no (verdi.nethelp.no [158.36.41.162])
	by hub.freebsd.org (Postfix) with SMTP id 61A3C14EE0
	for <freebsd-isp@FreeBSD.ORG>; Wed,  3 Mar 1999 01:41:42 -0800 (PST)
	(envelope-from sthaug@nethelp.no)
Received: (qmail 9381 invoked by uid 1001); 3 Mar 1999 09:41:24 +0000 (GMT)
To: kpielorz@tdx.co.uk
Cc: grant@vbc.net, freebsd-isp@FreeBSD.ORG
Subject: Re: ECC formatting
From: sthaug@nethelp.no
In-Reply-To: Your message of "Wed, 03 Mar 1999 09:35:31 +0000"
References: <36DD0263.58D2C566@tdx.co.uk>
X-Mailer: Mew version 1.05+ on Emacs 19.34.2
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Date: Wed, 03 Mar 1999 10:41:24 +0100
Message-ID: <9379.920454084@verdi.nethelp.no>
Sender: owner-freebsd-isp@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

> This sounds like it might be a hardware issue? (And has little to do with
> FreeBSD :-) - Are you sure the drives will do 528 bytes per-sector? - I'd
> imagine this would have to be done transparently?! - Do you really need this
> if the drives are in a RAID array anyway?

If you want hot-swap capability, you probably need it.

> Most (if not all) current drives have their own hardware / transparent ECC
> code on the drive, which uses things like read-solomon coding to detect and
> correct errors - and cannot be switched off (AFAIK).

The point of 528 bytes per sector is to get extra error detection/correction
info all the way from the platter to the RAID controller. The hardware ECC
on the drive isn't enough to cover electrical noise etc. on the SCSI bus
when you're doing hot swap.

Steinar Haug, Nethelp consulting, sthaug@nethelp.no


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message