Adding nodes to Cassandra Cluster

Question

I have a 1 node - 1 cluster Cassandra running on my local machine. I want to see how Cassandra scales up, when I simply add the second node to the same cluster. The second node that I am adding will also run from my local machine.

I have been trying to lookup a lot of approaches to do this, but can't find any. Any simple way through which I can add second instance of Cassandra to my cluster?

Gujral

Maybe the easiest (slow, though) would be to have a VPS that has a shared IP address so you can connect to/from it. I have done that to test with many nodes, but the hard drives are just really slow when you have too many in this way. – Decarbonize

Answer 1

The act of adding a second node is known as bootstrapping. The procedure is straightforward:

• Make sure the cluster names match in cassandra.yaml

• Point the new (joining) node to the cluster using the seeds setting in the yaml

• Start cassandra, it will join.

This is described in detail online, but one of the easiest ways to see it in action is using ccm , which is a tool many cassandra developers use to setup fake clusters on their local computers.

Currently (all versions up to 3.9), Cassandra's unique ID is an IP - you can only have one instance per IP. If you need more than one cassandra instance per computer, you need to use multiple IPs.

Otalgia

It is "straightforward" when you use one computer per node. – Decarbonize

so, there´s nothing that I´ve to do in the old node? – Allot

You should run "nodetool cleanup" on the old node after the new node has finished joining (shows UP/NORMAL in the ring). That will remove the data that was previously owned, but is no longer owned by that instance. – Otalgia