A cluster is a dynamic system in which nodes might be stopped either deliberately or by unforeseen circumstances, or nodes might be added to the cluster on-the-fly to scale the database for improved performance. The voltdb start command provides the following additional functions, described later in this book, for rejoining and adding nodes to a running VoltDB database:
Section 10.3, “Recovering from System Failures” — Use the same voltdb start command to start the cluster or rejoin a failed node.
Section 9.3.1, “Adding Nodes with Elastic Scaling” — Use voltdb start with the --add flag to add a new node to the running database cluster.
Documentation