advancedHigh Availability: Multi-Node Control Plane & Leader Election

Why are etcd clusters deployed with an odd number of nodes instead of an even number?

etcd requires a majority (quorum) of nodes to agree before a write is committed — an odd count (3 or 5) guarantees a network partition always leaves at most one side with a clear majority, while an even count (like 4) can split evenly with neither side able to reach quorum at all.

Ready to master this question?

Generate a complete walkthrough — background, the full answer in plain language, a working code example explained line by line, a real-world scenario, common mistakes, and how this same question gets asked in different ways.

Sign in to generate a response

Next Step

Continue to Does the Kubernetes API Server use leader election like the Scheduler does?← Back to all Kubernetes questions