Very Computer

Hello,

I am attempting to install Debian3.0r1 (2.4.18 kernel) onto a 2TB
3ware RAID array. During the partition portion of the installation
process, negative device size numbers appear. Breaking the array into
sizes less than 1 TB mitigates the issue.

I have searched for large block device limitations and found that 2TB
is the well-known 2.4 kernel limitation. Because I do not surpass this
limitation and seem to have problems with any device size greater than
1 TB, I do not believe I am running into a kernel problem (although,
maybe someone knows otherwise). It appears to be a signed vs. unsigned
integer issue, potentially in the partitioning utilities.

Has anyone seen and/or run into this issue before? I get the same
results during the partitioning process using cfdisk and fdisk.  A
Redhat 9.0 installation does not seem to have the same issue (although
people did report problems with earlier 2.4 kernels shipped with
Redhat).  Does Redhat use a kernel backport (from Peter Chubb or
something)?  Is this a known kernel issue, app issue, and/or Debian
issue?

Thanks much for any help,
Joe

Hi!

I have just a 1.6TB fs in test here... what fs type did you
choose?

CU Lars.

What are you going to be using the filesystem for? For most uses, I would
recommend XFS, as that brings many useful features from the commercial UNIX
world. See its website (http://oss.sgi.com/projects/xfs/) for details.

--

-- http://os2.dhs.org/~skylar/

Hi!

At the moment, we have everything on reiserfs (which is the fs of our
servers since 2000). However, I am currently installing two servers
connected to fibre channel storage, and I would prefer a shared fs there,
so I will try out gfs (we want to access the same fs from both servers).

CU Lars.