Oh, I presume this is the value of HADOOP_HEAPSIZE specified in the
hadoop-env.sh and is the heap given to the deamon by Java.
I guess this answers the question.
On Tue, Mar 2, 2010 at 10:46 PM, Saptarshi Guha
<[EMAIL PROTECTED]> wrote:
> I often notice "Cluster Summary (Heap Size is 281/889)" where the 281
> is much when I start up Hadoop.
> How does this relate to Hadoop Mapreduce?
> Thank you