It can be a problem with hazelcast protocol. Do you have Load Balancer behind these two servers? Could you please check table called pegadata.pr_sys_statusnodes. You should have there a list of nodes that working in cluster. If you do not see there all nodes it means that there is miss-configuration.