From owner-freebsd-questions  Fri May 17 16:19:45 1996
Return-Path: owner-questions
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id QAA28456
          for questions-outgoing; Fri, 17 May 1996 16:19:45 -0700 (PDT)
Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id QAA28404
          for <freebsd-questions@freefall.freebsd.org>; Fri, 17 May 1996 16:19:37 -0700 (PDT)
Received: (from root@localhost) by dyson.iquest.net (8.7.5/8.6.9) id SAA04661; Fri, 17 May 1996 18:19:00 -0500 (EST)
From: "John S. Dyson" <toor@dyson.iquest.net>
Message-Id: <199605172319.SAA04661@dyson.iquest.net>
Subject: Re: max filessystem size in FreeBSD 2.2?
To: terry@lambert.org (Terry Lambert)
Date: Fri, 17 May 1996 18:19:00 -0500 (EST)
Cc: kuku@gilberto.physik.rwth-aachen.de,
        freebsd-questions@freefall.freebsd.org
In-Reply-To: <199605172138.OAA20794@phaeton.artisoft.com> from "Terry Lambert" at May 17, 96 02:38:14 pm
X-Mailer: ELM [version 2.4 PL24 ME8]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-questions@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

> > What is the maximum filesystem size?
> 
> 2^63 or 9 * 10^18 bytes.  There is a disklabel issue at 128G;
> patches are available (yes, people are running with larger than 128G
> FS's).
> 
> The maximum single file size is on the order of 2^48 (~280G) because
> of VM limitations.  Ask John Dyson to be sure.
> 
> 
Current VM limitations are probably due to the PAGE_SIZE being 4K, and
we can have 32 bits of them.  Whether it is 2GPages or 4GPages depends
on my code signed/unsigned bogosity factor :-), or using the "right
variable sizes."

There is probably a limitation in vfs_bio (or other filesystem code) of
512Bytes for a block and we can have 32 bits of them.  That would
cause me to guess that at that layer we can probably do 1Tb (512*2GBlocks).
Again, there might be a problem with using the "right size" of variables.

There are other limitation that I have heard of, but those are the basic
limitations of the above subsystems.  It is possible to overcome them,
and I would think that the vfs_bio limitation is worse.

John