If HDFS is deployed with a single name node, what limitations may exist? (Multiple choices)

Question

Options

• AThe throughput of the entire distributed file system is limited by the throughput of a single name
• BIt will affect the function realization of the upper layer components based on HDFS
• CThe number of objects (files, blocks) that the name node can hold will be limited by the size of the
• DOnce the only-name node fails, the entire cluster will be unavailable

• A
  79% (34)
• B
  21% (9)

Explanation

In HDFS, the NameNode is the master metadata server that keeps all file system namespace information in memory and handles every client request - making it the central bottleneck. A is correct because every read/write operation requires a metadata lookup through the single NameNode, so its processing capacity caps the throughput of the entire cluster. C is correct because the NameNode stores all file and block metadata in heap memory (RAM), meaning the total number of objects the cluster can track is hard-limited by the physical memory of that one machine. D is correct because the NameNode is the sole holder of namespace state - if it goes down, no client can locate any data blocks, rendering the whole cluster inaccessible (classic single point of failure).

Why B is wrong: Upper-layer components like MapReduce, Hive, or HBase still function correctly - they just inherit the constraints from A, C, and D. A single NameNode constrains capacity and availability, not the functional correctness of higher-level components, so B overstates the impact.

Memory tip: Picture the NameNode as the only waiter in a restaurant - they can only serve so many tables at once (A), can only memorize so many orders (C), and if they quit, the whole restaurant shuts down (D) - but the kitchen (upper-layer components) still knows how to cook.

No community discussion yet for this question.