Fixing Node performance imbalance

I'm looking at an 8 node H500 cluster that has a severe performance imbalance.  Nodes 1 - 4 are consistently running at about 90% utilization, while nodes 5 - 8 are steady at 30% utilization.  The cluster was built as 4 nodes and then expanded after some time in use.  Client connections and protocol usage are level across all 8 nodes.  It looks like there is some key data that is resident on the first 4 nodes that was not evenly restriped when the new nodes were added, which is showing as a hot-spot on the initial nodes.  InsightIQ is installed and running, so I was wondering if there is standard procedure for identifying this data and forcing a restripe so it is more evenly distributed across all 8 nodes.  Any advice would be appreciated.