quorum
The minimum number of members, more than half, that must agree before a replicated system accepts a change.
In more detail
With three members the quorum is two, so one can fail; with five it is three, so two can fail. Even numbers add risk without adding tolerance, which is why members come in odd numbers. Quorum protects availability, not data: a deletion is replicated to every member, so it is not a backup.
Example
A three-member etcd cluster keeps accepting writes with one member down.Related terms
- etcd — The distributed key-value store where Kubernetes keeps all cluster data, including every Secret.
- Raft — An agreement method that lets several copies of a database accept writes in the same order and keep working when a member fails.
- high availability — Designing a system so it keeps working when one part fails, usually by running redundant copies with no single point of failure.
- etcd snapshot — A point-in-time copy of the whole etcd database, which is the backup of a Kubernetes cluster's state.
- CAP theorem — The result that during a network partition a replicated system must choose between consistent answers and answering every request.
- replication — Keeping copies of data in more than one place, updated automatically, so one failure does not lose or block it.
- strongly consistent read — A database read guaranteed to return the result of every write that finished before it, never an older copy.
Taught in
- etcd: Snapshot, Restore, And The Cache That Lies — Certified Kubernetes Administrator (CKA) Exam Prep
- HA Control Planes: Quorum, Stacked Or External — Certified Kubernetes Administrator (CKA) Exam Prep
- Control Plane: etcd, API Server, Scheduler, Controllers — Kubernetes: Production-Grade Container Orchestration
- StatefulSets, Stable Identity And Data Recovery — Kubernetes: Production-Grade Container Orchestration
- kubeadm, High Availability And Cluster Lifecycle — Kubernetes: Production-Grade Container Orchestration
- etcd Snapshots And A Tested Restore — Kubernetes: Production-Grade Container Orchestration
- Consistency, And The Truth About CAP — Scalable System Design & Distributed Architecture
References
- etcd: FAQ on cluster size and failure tolerance
- Apache Cassandra documentation: Dynamo (consistent hashing, replication, quorums)
- Patterns of Distributed Systems: Majority Quorum