Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Jun 2001 11:16:27 +0930
From:      Greg Lehey <grog@FreeBSD.org>
To:        "Forrest W. Christian" <forrestc@imach.com>
Cc:        freebsd-isp@freebsd.org
Subject:   Re: RAID/vinum options (mail server)
Message-ID:  <20010621111627.I86129@wantadilla.lemis.com>
In-Reply-To: <Pine.BSF.4.21.0106201203200.11185-100000@workhorse.iMach.com>; from forrestc@imach.com on Wed, Jun 20, 2001 at 12:15:23PM -0600
References:  <Pine.BSF.4.21.0106201203200.11185-100000@workhorse.iMach.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday, 20 June 2001 at 12:15:23 -0600, Forrest W. Christian wrote:
> I'm needing to find a RAID solution for our mail server.
>
> This is primarily for data integrity.   So the solution MUST be
> robust.  In the past I've had raid controllers which freaked at loosing a
> disk and refused to work with the array.  Needless to say I was pissed.
>
> I am not opposed to vinum.   I would like to hear more about it's
> restrictions specifically in regards to the boot partition (meaning can I
> vinum a slice on the same disk as the root partition) and it's robustness
> on disk failure.   The docs for vinum aren't too specific on this, unless
> someone knows of some docs I don't know about.

You currently can't have the root partition on a Vinum volume.  That's
the only limitation; you can have anything else, including swap, on a
Vinum partition.  Setting it up is a little tricky: you first need to
install normally, then redefine the partitions as Vinum volumes by
creating a Vinum drive (a partition of type Vinum) which starts 265
sectors before the beginning of the first partition, then defining
subdisks to match the position of the partitions relative to the start
of the drive.

Robustness is good.  You can have a failure and continue *if* you're
in a redundant configuration (RAID-1 or RAID-5).  You can also change
disks online if you have a hot-pluggable configuration.  We've seen
some data consistency problems in the RAID-5 code, mainly relating to
kernel race conditions.  I have fixed all known issues, and I haven't
heard of any problems since, but I certainly feel more confident about
RAID-1 or RAID-10, which hasn't had any consistency issues reported in
nearly 2 years.  RAID-1 will also give you much better write
performance than RAID-5 (on any platform; this is a limitation of the
technique, not Vinum).

Greg
--
See complete headers for address and phone numbers

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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010621111627.I86129>