Gluster users - Aug 2013 - Heavy workload on a specific node (replicate volume)

I have two nodes running gluster 3.3, and have created a replicate volume
(replication=2) on them. Then I use another server to mount the volume
using native driver and create 200 processes to continuously write small
files (10~15MB per file) to the volume.

The top command shows that one of the node has much higher workload.

My questions are:
1) Is it normal?
2) Is there anything I can do to balance the workload among these two nodes?
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://supercolony.gluster.org/pipermail/gluster-users/attachments/20130809/70c56a51/attachment.html>