Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 15 Mar 1998 17:15:06 -0800 (PST)
From:      Simon Shapiro <shimon@simon-shapiro.org>
To:        Raul Zighelboim <rzig@gulfsouth.verio.net>
Cc:        "scsi@freebsd.org" <scsi@FreeBSD.ORG>
Subject:   RE: to raid or not to raid
Message-ID:  <XFMail.980315171506.shimon@simon-shapiro.org>
In-Reply-To: <A03CD00C69B1D01195AB00A024ECEB167110DC@kaori.communique.net>

next in thread | previous in thread | raw e-mail | index | archive | help

On 16-Mar-98 Raul Zighelboim wrote:
> 
> My missing link is:
> 
> Will a (7 drive array under raid 5) be faster than a sinlg larger drive
> ?...
>        Will it be as fast as a (5 drive raid 5 array) ?

Depends a lot on who/what executes which RAID code.  If you are running
in-kernel RAID on a 486 which is overwhelmed with (let's say) network
traffic, you may suffer greatly.  If you use a deidicated RAID controller
that does all the RAID-5 bit twidling in dedicated hardware and has 100IPS
and 256MB of RAM to do it with, RAID-5 may appear faster than another
solution's RAID-0.

All else being equal, RAID-5, in optimal state, should be able to read
almost as fast as RAID-0. Should be able to write about as fast as a single
disk.

increasing the number of stipes (drives) in a RAID-5 arrays slows WRITE
operation worse than linear.  READ improves more or less linearly.

I have successfully run 10-12 drives in a RAID-5 arrangement on a DPT
PM3334UDW, balanced on ONE bus.  Throughput is about 8/4 MB/Sec READ/WRITE.
On a dual bus arrangement, 14 drives went up to 15/7.  Lots of caching.
But this is not an inexpensive arrangement.  Generally you can assume that,
for a Unix filesystem traffic, you will saturate aSCSI bus with 4-8 drives.

Simon


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



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.980315171506.shimon>