October 2018
Beginner to intermediate
348 pages
10h
English
It is important to remember that, once you have completed adding all of the new nodes, the data streamed to them still resides on other nodes. The reason for this is that, as new nodes get added, the number of tokens that each node is responsible for becomes smaller. The original nodes will still have the data that they have streamed but are no longer responsible for. In larger clusters, this data can be sizable.
To ensure that a node only contains data for which it is responsible, you can invoke the nodetool cleanup command:
nodetool cleanup