Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 5 Dec 1996 23:21:13 +0100
From:      se@freebsd.org (Stefan Esser)
To:        mrcpu@cdsnet.net (Jaye Mathisen)
Cc:        kuku@gilberto.physik.rwth-aachen.de (Christoph Kukulies), freebsd-hackers@freefall.freebsd.org
Subject:   Re: ccd considerations
Message-ID:  <Mutt.19961205232113.se@x14.mi.uni-koeln.de>
In-Reply-To: <Pine.NEB.3.95.961205134213.1925J-100000@mail.cdsnet.net>; from Jaye Mathisen on Dec 5, 1996 13:43:28 -0800
References:  <199612050844.JAA17324@gilberto.physik.rwth-aachen.de> <Pine.NEB.3.95.961205134213.1925J-100000@mail.cdsnet.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Dec 5, mrcpu@cdsnet.net (Jaye Mathisen) wrote:
> 
> I would think that you would have to have some pretty high load to require
> 2 controllers.

Not if the drives used have only very small caches.
Then you may have one drive's buffer overflow before
the other one disconnects for the first time ...
(The low end SCSI drives now come with only 128KB of
static RAM, some 30% of which are used by the drive's
firmware and are not available for data caching and 
read ahead. I just don't understand why Quantum 
bothered to support Tagged Command Queues on such a 
drive (the Fireball TM), with some 90KB of RAM to 
hold data recently read, to be written and possibly
read-ahead ...

Regards, STefan



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19961205232113.se>