Do Terracotta clusters have a single point of failure?
No. All data in a Terracotta cluster is managed by the Terracotta Server Array. The Server Array can be configured to run in a two-server active-standby configuration to provide failover services. The standby server immediately takes command if the active fails, while the active automatically becomes the passive when restored. Available in the Terracotta FX product is the ability to configure multiple server group stripes for added scale.