This can be caused by either having redundant relay servers in the configuration or the relay server running out of resources while creating the list of all relay servers. Make sure the configuration has uniquely-named relay servers and/or make sure there are sufficient resources on the machine running the relay server.