Confluent CCAAK - Confluent Certified Administrator for Apache Kafka

With min.insync.replicas=3 and acks=all, a partition must have at least 3 in-sync replicas (ISRs) for the producer to succeed. With one broker down (broker 0), only 2 ISRs remain → not enough to meet the requirement. Therefore, producers cannot write to affected partitions.

However, for some partitions where the leader and replicas are on brokers 1 and 2, the full ISR (3) may still be present (if broker 0 was not a replica). In this scenario, exactly 6 of the 9 partitions may still have ISR=3 depending on replica distribution.

From Kafka documentation:

“When acks=all and min.insync.replicas is not met, the broker will reject the produce request with an exception.”

Page Reference:

Kafka: The Definitive Guide, 1st Edition, Chapter 5, p. 159–161

Apache Kafka Documentation: Topic Config – min.insync.replicas

Kafka's Controller is a broker that also performs controller duties. It is not a dedicated node. Therefore, it continues to host topics, partitions, and leaders just like any other broker.

From Kafka documentation:

“The controller is one of the brokers in the cluster and continues to perform all normal broker functions.”

A → Correct: The controller is also a normal broker with partitions.

B/C/D → Incorrect: Replication and consumer fetches are not tied to the controller.

Page Reference:

Kafka: The Definitive Guide, 1st Edition, Chapter 6, p. 190

Apache Kafka Documentation: Kafka Controller Behavior

‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾

The kafka-consumer-groups.sh script is used to monitor consumer group offsets and lag. It shows the current offset, log end offset, and lag for each partition consumed by a consumer group.

From Kafka documentation:

“You can use kafka-consumer-groups.sh with the --describe flag to get the offset lag per consumer group.”

A → Incorrect: No such built-in script.

B → Correct: Official script for viewing consumer group lag.

C → Incorrect: Does not exist as a built-in Kafka tool.

D → Incorrect: Used for partition reassignment, unrelated to lag monitoring.

Page Reference:

Kafka: The Definitive Guide, 1st Edition, Chapter 10, p. 308

Apache Kafka Documentation: kafka-consumer-groups.sh usage

Unclean leader election allows Kafka to elect a non-in-sync replica as a new leader when all in-sync replicas are unavailable. This increases availability at the cost of possible message loss (sacrificing reliability).

From Kafka documentation:

“If unclean.leader.election.enable is set to true, a broker that is not in the ISR can be elected as leader, which increases availability but may lead to data loss.”

A → Correct: Improves availability when ISR is empty.

B → Best practice for quorum-based systems like ZooKeeper but not the setting that trades off reliability.

C → Controls batching latency, unrelated to availability.

D → Retention -1 means “retain forever,” which impacts storage, not HA.

Page Reference:

Kafka: The Definitive Guide, 1st Edition, Chapter 6, p. 195

Apache Kafka Documentation: Topic Config – unclean.leader.election.enable

‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾

Kafka supports topic deletion via the kafka-topics.sh tool if the broker property delete.topic.enable=true. This is the recommended and cleanest method to delete a topic and its data from the cluster.

From Kafka documentation:

“To delete a topic, use kafka-topics.sh with --delete flag, provided delete.topic.enable is set to true on the broker.”

A → Correct: Official method.

B → Misunderstood use of null keys (used for compaction).

C → Not recommended; causes inconsistent state.

D → Deletes consumer state, not the topic itself.

Page Reference:

Kafka: The Definitive Guide, 1st Edition, Chapter 6, p. 198

Apache Kafka Documentation: Topic Deletion

‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾

To ensure that duplicate messages are not processed and no messages are skipped, ksqlDB’s persistent queries must be configured to provide exactly-once processing guarantees. This is achieved by setting the processing.guarantee property to exactly_once. This configuration ensures that messages are neither lost nor processed more than once in the event of failures or restarts.

As per the official Confluent documentation:

“ksqlDB persistent queries can be configured for either at-least-once or exactly-once processing semantics by setting the processing.guarantee configuration property.”

“When processing.guarantee is set to exactly_once, ksqlDB ensures that messages are processed only once, even in failure scenarios. This eliminates duplicates and guarantees strong message delivery semantics.”

Other options:

Option B and C (ksql.streams.auto.offset.reset) only control the starting offset for new consumers, and do not enforce exactly-once semantics.

Option D (ksql.fail.on.production.error) controls whether errors during message production should fail the query, and does not affect exactly-once processing.

Page Reference:

Confluent Kafka: The Definitive Guide, 1st Edition, Chapter 11 (Stream Processing with ksqlDB), p. 365.

Confluent Documentation, “ksqlDB Configuration Properties” (processing.guarantee).

‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾