In which situations of distributed storage can it not cause reconstruction?

Question

Options

• AThe hard disk has not been restored after being offline for 10 minutes.
• BLong-term damage to the hard disk
• CThe hard disk is offline 5 times within 10 minutes
• DNormally shut down a node

• A
  11% (3)
• B
  7% (2)
• C
  4% (1)
• D
  78% (21)

Explanation

Normally shutting down a node is a planned, graceful operation where the distributed storage system receives a proper shutdown signal. The system knows the node will return, so it simply marks the node as temporarily unavailable and holds off on reconstruction - no data is considered lost.

Why the distractors are wrong:

• A - After an unplanned offline period exceeding a threshold (e.g., 10 minutes), the system assumes the disk may be permanently lost and triggers reconstruction to restore redundancy.
• B - Long-term hardware damage means data is genuinely at risk of being unrecoverable, so reconstruction is essential to maintain the required replica/erasure-coding count.
• C - A disk that flaps offline repeatedly in a short window is flagged as unstable/unreliable; the system treats it as a failing drive and initiates reconstruction to migrate data to healthy storage.

Memory tip: Think "planned = no panic." A normal shutdown is the only scenario where the system has advance notice and expects the node to return, so it waits rather than reconstructing. Every other option represents an unexpected or repeated failure, which forces the system to act defensively to protect data durability.

No community discussion yet for this question.