April 2018
Intermediate to advanced
468 pages
14h 34m
English
A StatefulSet is great, but, as mentioned earlier, Cassandra is already a sophisticated distributed database. It has a lot of mechanisms for automatically distributing, balancing, and replicating the data around the cluster. These mechanisms are not optimized for working with network persistent storage. Cassandra was designed to work with the data stored directly on the nodes. When a node dies, Cassandra can recover having redundant data stored on other nodes. Let's look at a different way to deploy Cassandra on a Kubernetes cluster, which is more aligned with Cassandra's semantics. Another benefit of this approach is that if you have an existing Kubernetes cluster; you don't have to ...
Read now
Unlock full access