Coupang Infra Interview: Kubernetes Storage Management

This question tests your practical understanding of Kubernetes persistent storage and how you manage stateful workloads in production.

You will be asked to explain key components—PersistentVolume (PV), PersistentVolumeClaim (PVC), StorageClass—and the lifecycle that connects them (provisioning, binding, using, releasing, reclaiming). Expect to walk through both static and dynamic provisioning: when you define PV objects yourself versus when the cluster provisions storage on demand using a StorageClass and a CSI driver.

The interview typically progresses in stages: first you define concepts and trade-offs (e.g., reclaim policies, access modes, ReadWriteOnce vs ReadWriteMany), then you describe a hands-on setup (YAML for PV/PVC, associating a claim with a Pod or StatefulSet), and finally you cover operational concerns (resizing PVCs, snapshots, backups, access control, multi-tenant storage isolation). You may be asked to sketch a storage class that favors performance (local SSD, IOPS) or durability (replicated block storage) and to explain how CSI fits into dynamic provisioning.

To demonstrate readiness, show familiarity with common use cases (databases, file shares, logs), explain trade-offs (performance vs durability, shared vs exclusive mounts), and mention tooling/commands you’d use (kubectl, CSI snapshotter, storage provisioner logs). Practical examples and failure-recovery steps (reclaim policy handling, restoring from volume snapshots) strengthen your answer.