Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 28 Jun 1998 13:01:08 +0930
From:      Greg Lehey <grog@lemis.com>
To:        Richard Foulk <richard@pegasus.com>, Stefanos Kiakas <stefanos@ringworld.uniscape.com>, freebsd-hardware@FreeBSD.ORG
Subject:   Re: RAID and FreeBSD
Message-ID:  <19980628130108.I28055@freebie.lemis.com>
In-Reply-To: <199806280325.RAA25541@pegasus.com>; from Richard Foulk on Sat, Jun 27, 1998 at 05:24:53PM -1001
References:  <grog@lemis.com> <199806280325.RAA25541@pegasus.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday, 27 June 1998 at 17:24:53 -1001, Richard Foulk wrote:
>>> 	I have been using FreeBSD for the last year and I am interested
>>> in setting up a RAID server. I was wondering what hardware and what software
>>> choices I have under FreeBSD? Specifically what hardware is support (works),
>>> and what is in beta.
>>>
>>> 	I know about ccd, but I'm looking for other solutions as well. I would
>>> appreciate any help you can provide.
>>
>> Check out http://www.lemis.com/vinum.html.  It's a replacement for ccd.
>
> Having had a software RAID system fail I would recommend against them.

What would you have done if you had had a hardware RAID system fail?

> RAID is supposed to add reliability.  Done in software in the kernel it
> adds complexity and uncertainty.

RAID anywhere is a complex business.  And yes, it has bugs.  We've
heard, for example, of a number of bugs in the DPT (hardware) raid
boxes.

> With hardware RAID there is greater assurance that once you've used
> it for a few weeks or months that it will remain consistent.  As
> part of an OS kernel (drivers, etc.) uncertainty revisits every time
> the system software is upgraded, patched or changed in some way.

I suppose if you don't upgrade your RAID box, you don't change old
bugs for new.  The same applies to a kernel, of course.

> Hardware RAID may be a touch slower, but it's more trustworthy.

I'm not contesting this, though I'm surprised you think it's slower.
I would expect it to be faster, since you have another processor to
which you can offload the work.  At the moment, of course, we don't
have anything to go on.

The real question is: how can each fail?  How great is the likelihood
of losing data?  ccd is a particularly poor choice in this matter,
since even with mirroring, the failure of one mirror takes it down and
you have to reconfigure.  vinum doesn't have this problem.

I'd like to discuss this matter, since it could have some effect on
how vinum develops.

Greg
--
See complete headers for address and phone numbers
finger grog@lemis.com for PGP public key

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



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980628130108.I28055>