It looks like there was a heap space issue, you may need to increase the heap size.
Can the nodes communicate with each other across the hazelcast ports? (5701-5709 in your case it looks like)
Do you have multiple Network Interface Cards on the server hosting the nodes? If so, you may need to specify the IP that should be used for clustering in the prconfig.xml file using the following setting:
<env name="cluster/hazelcast/interface" value="xx.xx.xx.xx"/>