[Ocfs2-users] Different size with du and ls

Sunil Mushran Sunil.Mushran at oracle.com
Thu Jul 10 14:44:13 PDT 2008


Markus Meyer wrote:
> 	Block Size Bits: 12   Cluster Size Bits: 16
> 	Links: 0   Clusters: 6707596

So you have a 4TB volume. Correct? Appears mkfs chose 64K as the
cluster size. This means the smallest data allocation would be 64K.

>    File: `/mnt/user/small/11/11zzzzwa1.jpg'
>    Size: 2420      	Blocks: 128        IO Block: 65536  regular file

See how while the filesize is 2420 bytes, but the ondisk allocation is 
128 blocks.
stat uses 512 byte blocksize making it 64K.

>    File: `/data/user/small/11/11zzzzwa1.jpg'
>    Size: 2420      	Blocks: 8          IO Block: 4096   regular fil
xfs consumes 8 512byte blocks or 4K.

So, du is not wrong. By default it lists the ondisk usage which is 
higher for
ocfs2.

One solution is to format ocfs2 with explicit values.

mkfs.ocfs2 -b4K -C 4K -L label -T mail -N 2 /dev/sdX
This should force it to use 4K clustersize.

Sunil



More information about the Ocfs2-users mailing list