SOLVED [rhn-users] file size problems

Chuck Campbell campbell at accelinc.com
Mon Mar 7 21:22:32 UTC 2005


The solution is to force nfs version 3, which has large file support, and 
all is well.

-chuck


On Mon, Mar 07, 2005 at 06:29:26AM -0600, Chuck Campbell wrote:
> On Sun, Mar 06, 2005 at 04:57:30PM -0800, Geoff Sweet wrote:
> > Well ext3 certainly has large file support.  I am trying to rack my
> > brain if there is a max filesize limitation that can be set either from
> > a user standpoint, or from an NFS standpoint.  Can you try creating a
> > large file with something like
> > 
> > dd if=/dev/zero of=/NFS_MOUNT/bob.huge bs=1024k count=3000
> > 
> > on the NFS share, and then again on the local machine?  That might help
> > determine if maybe NFS is limiting the file-size.  Just a thought for
> > trouble shooting
> 
> This also fails at two GB (2147483647 bytes actually).
> 
> coreutils is at version 4.5.3.  Is there something newer?  Do I need to turn 
> on large file support explicitly?  If so, how?
> 
> I tried:
> dd if=/dev/zero of=/local_disk/big_file.tst bs=1024k count=3000
> and it ran without error.
> 
> I also ran this:
> dd if=/dev/zero of=/local_disk/big_file.tst bs=1024k count=3000
> without failure as well.
> 
> I'm at a loss, because the ftp of 33GB succeeded, this must be an 
> NFS limitation?  How do I fix this?
> 
> -chuck
> 
> > 
> > -Geoff
> > 
> > -----Original Message-----
> > From: rhn-users-bounces at redhat.com [mailto:rhn-users-bounces at redhat.com]
> > On Behalf Of Chuck Campbell
> > Sent: Sunday, March 06, 2005 9:57 AM
> > To: rhn-users at redhat.com; taroon-list at redhat.com
> > Subject: [rhn-users] file size problems
> > 
> > I just built two dual Opteron machines for a small cluster, with 4 GB of
> > ram 
> > each, a dedicated GB network layer 3 switch and about a TB of disk.
> > 
> > I put RHEL3 update 2, WS on the compute node and ES on the master.  All 
> > filesystems are ext3.
> > 
> > I loaded a 33GB file vi a ftp from another machine onto the cluster, and
> > 
> > had no issues, but when I did a cp from the master node local disk to
> > the
> > slave node local disk (mounted via NFS), I only got 2 GB of data moved,
> > and
> > an error message saying "File size exceeded".
> > 
> > What am I doing wrong?  This is a 64 bit OS on 64 bit processors.  Where
> > did 
> > this 3 GB file limit come from?  Is it in cp, NFS, ????
> > 
> > any help is appreciated.
> > 
> > thanks,
> > -chuck




More information about the rhn-users mailing list