For files stored in N (N>1) copies, the HDFS client initiates a file read request. If the read copy node fails, it will directly return to failure and will not go to other copy nodes for reading.

Question

Options

• ATrue
• BFalse

• A
  16% (9)
• B
  84% (48)

Explanation

B is correct because HDFS is designed with fault tolerance at its core - when a client's read request to a particular replica node fails (e.g., due to a DataNode crash or network issue), the HDFS client automatically marks that node as failed and retries the read against one of the remaining N-1 replica copies, transparent to the application. This failover behavior is a fundamental guarantee of HDFS and is why replication exists in the first place.

Why A is wrong: Claiming the client immediately returns failure ignores HDFS's built-in replica-aware retry logic; a single node failure is precisely the scenario that replication is designed to survive, not surface as an error.

Memory tip: Think of HDFS replicas like a library with multiple copies of the same book - if one copy is checked out (or the shelf is blocked), the librarian doesn't send you home empty-handed; they simply direct you to another shelf with the same book. "N copies = N chances to read."

No community discussion yet for this question.