Problem running out of space by copying files to identical filesystems

UNIX HP-UX
1

I am trying to copy a filesystem from one server to another using rsync over the WAN. As far as I can tell, the two filesystems are identical but for some reason I cannot copy the last file because I keep running out of space.

SERVER 1:
mkfs -m <lvol>
mkfs -F vxfs -o ninode=unlimited,bsize=8192,version=4,inosize=256,logsize=256,largefiles /dev/vgprod/store15 2147188736
bdf <lvol>
Filesystem kbytes used avail %used Mounted on
/dev/vgprod/store15
2147188736 2146853928 332200 100% /data/store15

SERVER 2:
mkfs -m <lvol>
mkfs -F vxfs -o ninode=unlimited,bsize=8192,version=4,inosize=256,logsize=256,largefiles /dev/vgprod/store15 2147188736
bdf <lvol>
Filesystem kbytes used avail %used Mounted on
/dev/vgprod/store15
2147188736 2113312464 33611672 98% /data/store15

All of the files are about 34GB each. The very last file is 34,346,224,640 bytes and, as you can see, I only have 33.6GB available so the copy fails on the very last file. Why do these files fit perfectly into SERVER 1 but not into SERVER 2 if the OS version is the same, the JFS version is the same and all of the parameters appear to be the same? There must be something I am missing but I cannot figure it out.

I have also checked that acctresume and acctsuspend are the same on both servers as well.

2

Let's see the rsync command too. I suspect the problem is hard-links and/or soft-links. If you don't properly deal with hard-links, then you end up with two or more copies to some very big files. You also have to be careful not to cross the /proc file system.