Cassandra provides various configuration options both at column family level and system level; this lets us customize Cassandra's features to suit our requirements. Let's discuss these configuration options.
From our discussion about caches in the previous chapters, we know that Cassandra has two caching options:
Key cache: This caches partition keys and their respective data offset in the SSTable. When a key cache is hit, we'll save one disk seek for that SSTable search.
Row cache: This stores the entire row of a partition key; in the new version, we can specify the number of rows to be stored per partition key. This caches two disk seeks as the entire row is in the memory.
Both of these caches can be configured per column family using the CQL command as we discussed in Chapter 4, Read and Write – Behind the Scenes.
However, there are configuration options for both of these caching types in the Cassandra.yaml file, which lets us do configurations...