I'm having a problem where df and du report different disk usage on a raid10 array created with mdadm. Here is some output:

> du -sh /data01
1.2T /data01

> df -h /data01
Filesystem Size Used Avail Use% Mounted on
/dev/md0 1.4T 169G 1.2T 13% /data01

> df -i /data01
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/md0 91578368 -188528 91766896 0% /data01

> mdadm --detail /dev/md0
/dev/md0:
Version : 00.90
Creation Time : Wed Aug 20 00:14:12 2008
Raid Level : raid10
Array Size : 1465148928 (1397.27 GiB 1500.31 GB)
Used Dev Size : 732574464 (698.64 GiB 750.16 GB)
Raid Devices : 4
Total Devices : 4
Preferred Minor : 0
Persistence : Superblock is persistent

Update Time : Wed May 20 22:17:14 2009
State : clean
Active Devices : 4
Working Devices : 4
Failed Devices : 0
Spare Devices : 0

Layout : near=2, far=1
Chunk Size : 64K

UUID : 967aec6e:cbbe7390:aba2e195:aa6694ad
Events : 0.2059734

Number Major Minor RaidDevice State
0 8 96 0 active sync /dev/sdg
1 8 80 1 active sync /dev/sdf
2 8 48 2 active sync /dev/sdd
3 8 64 3 active sync /dev/sde

The 1.2T used from du and 1.2T free from df is a coincidence, since it has reported different numbers in the past. But the 1.2T used is the correct number.

I have a few other disks, including a hardware raid1 array, and df reports the correct usage for those. Any ideas on why this might be? The negative inode listing for IUsed seems to be the biggest hint towards a solution, but I can't figure out why it's saying that.

Google brings up some results, but they don't seem to be related to this problem.