All Questions
Consistent Hashing hard

A cluster runs at 80% capacity. Node A fails. Why is a standard Ring with Neighbors-as-Replicas bad here?

  1. A The entire load of Node A (10% of total) shifts entirely to Node B (its successor), potentially causing a cascade of failures.
  2. B The data becomes permanently inaccessible because a standard ring walk does not support decentralized replica recovery.
  3. C The cluster latency increases slightly as the remaining nodes must perform expensive floating-point math to re-calculate the ring.
  4. D It is not technically 'bad' because consistent hashing is mathematically guaranteed to maintain a perfectly balanced load at all times.

One system design question a day, with spaced repetition so the ones you miss come back.

Get the app

Related questions