Kafka - Partition

About

Data Partitions (Clustering of data) in Kafka

Each partition is an:

• ordered,
• immutable sequence of records

that is continually appended to—a structured commit log.

The records in the partitions are each assigned a sequential id number called the offset that uniquely identifies each record within the partition.

Each partition in the Kafka cluster has:

• a leader
• and a set of replicas among the brokers.

Articles Related

Purpose

Partitions:

• allow the log to scale beyond a size that will fit on a single server. 1 Kafka partition = 1 disk physical. Each individual partition must fit on the servers that host it, but a topic may have many partitions so it can handle an arbitrary amount of data.
• act as the unit of parallelism

* 1 Partition = 1 process
  * 2 partition = 2 process
  * ...

Structure

Topic 
   * -> partition 1
       * -> segment 11
       * -> segment 12
   * -> partition 2
       * -> segment 21
       * -> segment 22
.......

where:

• topic
• segment

Physic Server Distribution

The partitions of the log are distributed over the servers in the Kafka cluster with each server handling data and requests for a share of the partitions.

Record Partition Assignment

The producer is responsible for choosing which record to assign to which partition within the topic. See Kafka - Producer (Write / Input)

Replication

Each partition is replicated across a configurable number of servers for fault tolerance.

Replicated partitions are called replicas

Leader / Follower

The leader handles all read and write requests for the partition while the followers passively replicate the leader. If the leader fails, one of the followers will automatically become the new leader.

Each server acts as a leader for some of its partitions and a follower for others so load is well balanced within the cluster.

Each partition has one server which acts as the leader and zero or more servers which act as followers.

Rebalance

If you're running out of capacity, you will add more brokers to your cluster, and then rebalance the partitions in your cluster to move the load around.

You can do it using

• the kafka-reassign-partitions script,
• Confluent Auto Data Balancer
• Cruise Control
• kafka-assigner