4. Storage

Distributed File Systems — Quiz

Test your understanding of distributed file systems with 5 practice questions.

Practice Questions

Question 1

In a lease-based cache coherence protocol in a distributed file system, what event typically triggers the server to revoke a client’s write lease on a file region?

Question 2

A file is striped across 8 storage nodes in a parallel distributed file system. Each node can sustain 125 MB/s of sequential read throughput. What is the maximum theoretical aggregate read throughput?

Question 3

In an erasure-coded distributed file system using a Reed–Solomon scheme with $k=6$ data fragments and $m=3$ parity fragments, what is the storage overhead and how many fragment failures can be tolerated without data loss?

Question 4

To optimize parallel I/O performance in HPC workloads, which scheduling strategy should compute frameworks employ when placing tasks relative to data in a distributed file system?

Question 5

Which metadata management architecture is most effective for scaling metadata performance under heavy shared workloads in a distributed file system?