Subject: Re: Formatting/Block sizes
To: Mark de Jong <mdj@home.com>
From: Dr. Bill Studenmund <wrstuden@loki.stanford.edu>
List: port-mac68k
Date: 09/03/1998 16:04:53
On Wed, 2 Sep 1998, Mark de Jong wrote:

> While using Mkfs, I noticed that it is allocating space as follows:
> 
> 	512 bytes / sector
> 	8192 bytes / block
> 	4096 bytes / inode
> 
> This appears to be an inefficient way of setting this up. It seems like a
> smaller block size would allow for more files to be stored on the disk.

Frederick Bruckman pointed out where to find in-depth details on what's
going on. But the simple answer is that an ffs file system can break each
block up into eighths. So with 8k blocks, the file fragment size is 1k.
That's a good number overall. If you were doing something like running a
news server, you might want 4k blocks, with a fragment size of 512 bytes.

But either way, it's better than HFS partitions, where the fragment size
is only 1k for SMALL partitions (65 MB or so ones).

Take care,

Bill