ResourceManager adopts a high-availability solution. When the Active resourcemanager finds a fault, it can only start the standby resourcemanager through the built-in zookeeper and switch its state…

Question

Options

• ATrue
• BFalse

• A
  17% (11)
• B
  83% (53)

Explanation

Option B is correct because the statement contains two inaccuracies about YARN ResourceManager HA. First, ZooKeeper is not built into ResourceManager - it is an external, independently deployed service that ResourceManager integrates with for leader election and state coordination. Second, ZooKeeper-based automatic failover is not the only method: administrators can also manually trigger a failover using yarn rmadmin -transitionToActive, making the word "only" factually wrong.

Option A is wrong because accepting this statement would incorrectly imply ZooKeeper is an embedded component of ResourceManager and that no manual failover path exists - both of which contradict how YARN HA actually works.

Memory tip: Remember the two red-flag words in the statement - "built-in" and "only." ZooKeeper is always external (never built-in to any Hadoop component), and RM failover is always flexible (automatic via ZooKeeper or manual via rmadmin). If an exam question uses absolutes like "only" in a Hadoop HA context, treat it with suspicion.

No community discussion yet for this question.