Which of the following are correct about the Name Node concurrency characteristics of Huawei's distributed storage products? (Multiple choice)

Question

Options

• AAvoid metadata service interruption caused by HDFS Name node switching
• BMulti-directory metadata operations on multiple nodes concurrently
• CMultiple Active NameNodes provide metadata services, and data is consistent across multiple
• DThe number of files supported by multiple Active Namenodes is no longer limited by the memory

• A
  79% (22)
• C
  7% (2)
• D
  14% (4)

Explanation

A and B are correct because Huawei's distributed storage implements a concurrent multi-NameNode architecture that eliminates the single point of failure inherent in standard HDFS: the design ensures seamless NameNode failover with no metadata service interruption (A), and partitions the namespace so different directory trees can be processed by different NameNodes simultaneously (B).

C is wrong because in HDFS federation-style architectures, multiple Active NameNodes each manage independent namespace partitions - they do not replicate or synchronize metadata with each other, so "data is consistent across multiple" misrepresents the model; consistency within each NameNode's namespace is maintained, but there is no cross-NameNode metadata consistency guarantee.

D is wrong because while spreading the namespace across multiple NameNodes distributes the memory load, each individual NameNode still stores its portion of the namespace in memory - the memory constraint is not eliminated, only partitioned; saying it is "no longer limited by memory" is too absolute.

Memory tip: Think "A = Availability (no failover gap), B = Balanced concurrency across nodes" - both describe how the system works. C and D fail because they make sweeping claims (cross-node consistency, zero memory limits) that go beyond what federation actually guarantees; when an option overpromises, it's usually a distractor.

No community discussion yet for this question.