End-to-End Kubernetes with Rancher, RKE2, K3s, Fleet, Longhorn, and NeuVector

Data Reliability with Replication

Replication in Longhorn is a key mechanism to ensure data reliability and guarantee high availability across our cluster. Each replica of a Longhorn volume stores a chain of snapshots that collectively represent the current state of the volume. These snapshots function like image layers, where the base layer is the oldest snapshot, and newer snapshots stack on top, incorporating only the data that has changed from earlier snapshots. Since unchanged data remains in older snapshots, this approach makes storage utilization efficient since modifications are captured incrementally in new snapshots.

For a Longhorn volume, multiple replicas are distributed across different nodes in the cluster, with each replica treated equally. When we created a Longhorn volume earlier, Longhorn automatically provisioned three replicas on our 3 worker nodes to ensure data redundancy and fault tolerance.

# SSH into the RKE2 CP
ssh root@$WORKLOAD_CONTROLPLANE_01_PUBLIC_IP

# List the Longhorn replicas
kubectl -n longhorn-system get replicas.longhorn.io