On Jul 7, 2010, at 6:48 AM, Michael Segel wrote:

> 
> Non DFS used tends to be logging or some other information on the disk.


It can also be the result of hadoop not properly parsing df.  [ZFS pools, for 
example.]

Reply via email to