Once .100 is up and running, I can add a second node to the cluster quite easily. We will assume that I now want to add 192.168.0.100 as a new node in my cluster.

All of the configuration properties mentioned prior to all of the files (cassandra.yaml, cassandra-env.sh, cassandra-rackdc.properties, jvm.options) should be specified exactly the same as the first node. The exceptions are these two lines in the cassandra.yaml file:

listen_address: 192.168.0.101
rpc_address: 192.168.0.101

Start the new node up, and tail the log to make sure it joins the cluster.

tail -f /var/log/cassandra/system.log

The seed node will stay the same for 101's configuration:

seeds: "192.168.0.100"

The seed node designation in Cassandra doesn't make a node any different. The seed list simply specifies nodes that should be contacted to gain information about the cluster when joining it. It is usually a good idea to have one or two seed nodes per logical data center.

The cluster_name of a new node must match the cluster_name of the target cluster that it aims to join.

While the node is joining, running a nodetool status should yield similar output to this:

$ nodetool status
Datacenter: LakesidePark
========================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address        Load       Tokens  Owns    Host ID                          Rack
UJ 192.168.0.101  83.99 MiB  16      100.0%  fd352577-6be5-4d93-8251-15a74... r40