In a Kafka cluster, when the broker running the controller thread fails, which broker will become the new controller?

Question

Options

• AThe broker having the next highest broker.id value
• BThe next broker to successfully recreate the Zookeeper ephemeral node
• CThe next broker to persist new metadata in Zookeeper
• DThe next broker in the cluster joined order

• A
  17% (4)
• B
  71% (17)
• C
  4% (1)
• D
  8% (2)

Explanation

In a Kafka cluster, the controller is the broker responsible for managing partition leadership and other cluster-wide operations. The controller is elected through a Zookeeper-based mechanism. When the current controller broker fails, the cluster needs to elect a new controller. The process works as follows: Each broker in the Kafka cluster creates an ephemeral node in Zookeeper (under /controller). If the current controller broker fails, the Zookeeper ephemeral node corresponding to the controller broker will be deleted. The broker that successfully re-creates the Zookeeper ephemeral node (after detecting the deletion) becomes the new controller. Thus, option B is the correct answer because the new controller is the broker that successfully re- creates this ephemeral node in Zookeeper.

No community discussion yet for this question.