Cassandra has no single point of failure or master nodes. Instead, it uses an internal process called Gossip to communicate changes about the topology of the Ring between nodes. Seeds are a list that a node attempts to contact on startup to begin gossiping. This recipe shows how to define seed nodes.
The technique for defining Seeds can vary based on the type and size of your deployment. This recipe describes the process for simple deployments using org.apache.Cassandra.locator.SimpleStrategy
.
For multiple datacenter deployments, place one or more seed in each datacenter, refer to Chapter 8, Multiple Datacenter Deployments.