🔒 Paid

Quorum Loss

A 5-node consensus group requires a genuine majority of its voting membership to agree before any write is considered safely committed. A network issue has isolated two of the five nodes, leaving the group at the exact edge of quorum with zero margin to lose another. If a third node also becomes unreachable, the group will genuinely fall below the number of nodes needed for a safe majority — and continuing to accept writes in that state is a correctness violation, not a capacity problem, regardless of how healthy write availability looks. Decide how to respond to a group under real threat of losing safe quorum, and what the right behavior is if that threshold is actually crossed.

Difficulty
Staff
Expected Time
~20 min
Category
Distributed Systems

Skills

  • Distinguishing a quorum genuinely threatened from one genuinely lost
  • Recognizing when customer-facing availability conceals an unsafe write
  • Reasoning about voting membership reconfiguration versus waiting for node recovery
  • Choosing a safe degrade (read-only) over an unsafe shortcut under node loss

System Architecture

  • Node A
  • Node B
  • Node C
  • Node D
  • Node E
  • Consensus Group

This is a Pro scenario. Upgrade to Pro to unlock it.

Upgrade to ProBack to Incidents