Interview Questions and Answers

The CAP theorem states that a distributed system cannot simultaneously provide more than two out of three guarantees: Consistency, Availability, and Partition Tolerance. Cassandra prioritizes Availability and Partition Tolerance over Consistency, making it an AP system.

A partition key is a primary key assigned to each row in a Cassandra table. It is responsible for distributing data across nodes in the cluster and is crucial for the performance of queries.

The Snitch in Cassandra is responsible for determining the proximity of nodes in a cluster. It helps in optimizing data distribution and ensures that data is stored on nodes that are geographically closer to each other.

The commit log in Cassandra is crucial for durability and fault tolerance. It stores write operations before they are written to the actual data files, ensuring that data is not lost in the event of a node failure.

Compaction in Cassandra is the process of merging and compacting SSTables (sorted string tables) to optimize storage and improve read performance.

Cassandra uses a write-ahead log (WAL) and a memtable for write operations. Data is first written to the commit log for durability and then stored in the memtable, which is periodically flushed to an SSTable on disk.

Eventual consistency in Cassandra means that, given enough time and in the absence of further updates, all replicas of a piece of data will converge to the same value. It allows for high availability and partition tolerance but may result in temporarily inconsistent data.

The Gossip Protocol is used by nodes in a Cassandra cluster to communicate with each other and share information about the state of the cluster. It helps in maintaining a decentralized and dynamic view of the cluster.

Cassandra uses a combination of partition key and clustering key to locate and retrieve data efficiently. Read operations can be served from memory (memtable) or disk (SSTables), depending on the specific scenario.

A wide row in Cassandra contains a large number of columns, while a narrow row has a smaller number of columns. The distinction is important for designing data models based on query requirements and performance considerations.

The coordinator node in Cassandra is responsible for receiving and coordinating client requests. It determines the nodes that need to be involved in the request and communicates with them to fulfill the operation.

A quorum in Cassandra is a majority of replicas that must respond for a read or write operation to be considered successful. The quorum level is configurable and is used to ensure consistency in distributed systems.