In Huawei Fusioninsight HD, which of the following components are strongly dependent on Flink? (Multiple choice)

Question

Options

• AKafka
• BHDFS
• CYARN
• DZookeeper

• A
  16% (4)
• B
  84% (21)

Explanation

In FusionInsight HD, Flink relies on HDFS (B) to persist checkpoints and savepoints for fault tolerance - without a distributed filesystem, Flink cannot durably store state. YARN (C) serves as Flink's resource manager, allocating containers and scheduling TaskManagers, making it a prerequisite for running Flink jobs in a cluster. ZooKeeper (D) is required for Flink's high-availability mode, handling JobManager leader election and storing job metadata so recovery is possible after failures.

Kafka (A) is a distractor - while Kafka is a popular source/sink connector for Flink streaming jobs, it is an optional integration. Flink can ingest data from many sources (files, sockets, custom connectors) and has no hard dependency on Kafka to function.

Memory tip: Think of B, C, D as Flink's infrastructure layer - storage (HDFS), resources (YARN), and coordination (ZooKeeper) - all three must exist before Flink can even start. Kafka sits above Flink as a data producer, not below it as a platform dependency.

No community discussion yet for this question.