H13-711_V3.5 · Question #248
Which of the following components must depend on Zookeeper to run? (Multiple choice)
The correct answer is A. HDFS B. HBase. HBase (B) has a hard dependency on Zookeeper - it uses Zookeeper for master election, RegionServer tracking, and client routing, and simply cannot start without it (HBase even ships with an embedded Zookeeper for convenience). HDFS (A) requires Zookeeper when configured with…
Question
Which of the following components must depend on Zookeeper to run? (Multiple choice)
Options
- AHDFS
- BHBase
- CSpark
- DYARN
How the community answered
(61 responses)- A77% (47)
- C8% (5)
- D15% (9)
Explanation
HBase (B) has a hard dependency on Zookeeper - it uses Zookeeper for master election, RegionServer tracking, and client routing, and simply cannot start without it (HBase even ships with an embedded Zookeeper for convenience). HDFS (A) requires Zookeeper when configured with NameNode High Availability (the standard production setup), where Zookeeper Failover Controllers (ZKFCs) coordinate automatic failover between Active and Standby NameNodes.
YARN (D) is wrong because while its ResourceManager can optionally use Zookeeper for HA failover, YARN runs fine without it - the HA configuration is optional, not mandatory. Spark (C) is wrong because Spark has no inherent Zookeeper dependency; it runs on YARN, Kubernetes, Mesos, or standalone without requiring Zookeeper (standalone HA can use Zookeeper, but that's optional).
Memory tip: Think "HB AZ" - HBase Always needs Zookeeper, and HDFS needs it for Active/standby ZKFC coordination. Services that merely can use Zookeeper (YARN, Spark standalone HA) are distractors - the key word is must.
Community Discussion
No community discussion yet for this question.