Subject: Re: newfs can't make filesystems over 1TB in size
To: Jaromir Dolecek <firstname.lastname@example.org>
From: Brian Buhrow <email@example.com>
Date: 12/06/2002 10:14:54
Hello Jaromir. According to the document you reference, the fragment
size must only be 2K to create a 4TB filesystem. The filesystem I'm trying
to create has a fragment size of 4K, and a block size of 16K. The problem
seems to be in newfs itself. The variable llsize is taken to be the size
of the filesystem, and llsize is a long long. However, the filesystem size
is passed to mkfs() inside fssize, which is a signed 32-bit integer. I
think here-in lies the problem.
After further investigation, and attempting to build a special version
of newfs without the signed integer restriction, I find that there are a
lot of assumptions that a filesystem will never be more than MAX_INT
sectors in size. I wonder if someone on this list has actually created a
filesystem over 1TB in size, and, if they have, if they'd care to share how
they did it?
The FreeBSD folks appear to have gotten around this problem by
defining fssize as a quad_t. How hard would it be to make the same change
Does anyone have experience trying to make filesystems of this size?
Does anyone mind trying to help me get around this limitation?