← Current Platform / runtime

Replicated block storage

Longhorn

Kubernetes-native replicated volumes for workloads that need node-independent persistent state.

CSIReplicated volumesSnapshotsk3s

01 / Why it exists

Some workloads need persistent volumes that can move with them without placing every stateful concern on shared NFS.

02 / Architecture

Longhorn replicates block volumes across eligible cluster nodes. It sits beside NFS and local-path storage rather than pretending one storage class fits every workload.

03 / Lessons

What operation taught me.

  1. 01Replication is not backup.
  2. 02Storage classes should express failure-domain intent.

Screenshots

Operational captures will be added when they can be safely published.

Source / Config

Public configuration will appear here when the boundary is ready.

Related material

Read the storage operating notes →