H13-723_V2.0 · Question #3
When the HDFS client writes files to HDFS with N copies, if one of the copies fails to write, all copies will return the write failure.
The correct answer is B. False. B is correct because HDFS uses a pipeline replication model with a minimum replication threshold - by default, a write is considered successful as long as at least dfs.replication.min copies (default: 1) are written successfully, even if some replicas fail. The client receives…
Question
When the HDFS client writes files to HDFS with N copies, if one of the copies fails to write, all copies will return the write failure.
Options
- ATrue
- BFalse
How the community answered
(36 responses)- A17% (6)
- B83% (30)
Explanation
B is correct because HDFS uses a pipeline replication model with a minimum replication threshold - by default, a write is considered successful as long as at least dfs.replication.min copies (default: 1) are written successfully, even if some replicas fail. The client receives a success acknowledgment once the minimum threshold is met, and the NameNode schedules re-replication of the missing copies asynchronously in the background.
Why A is wrong: HDFS is designed for fault tolerance. Requiring all N replicas to succeed before acknowledging a write would make the system fragile and defeat the purpose of replication - a single node hiccup would block every write operation.
Memory tip: Think of HDFS like mailing N copies of a document - if one gets lost in transit, you don't throw away all the others and start over. You just re-send the lost one. HDFS similarly recovers the missing replica without forcing the entire write to fail.
Topics
Community Discussion
No community discussion yet for this question.