Choose storage by semantics, not by product name

Block, shared file and object-backed mounts behave differently under failover, concurrency and backup. The application contract comes first.

One writer or many?

Persistent Disk is block storage; Filestore provides shared NFS access.

What performance changes?

Hyperdisk separates capacity from tunable IOPS and throughput.

Is it really a filesystem?

Cloud Storage FUSE exposes object storage with filesystem-like access.

What must be restored?

Backup for GKE protects workloads and volumes, not the cluster infrastructure itself.

Choose storage by semantics, not by product name

Storage orientation is easiest to understand by separating the Kubernetes contract from the Google Cloud implementation. Kubernetes CSI makes storage consumable through StorageClasses and PVCs, but it does not make different storage semantics interchangeable. The Kubernetes objects stay familiar, but GKE supplies controllers, infrastructure and safe defaults around them. This is why a team can move from an on-premises cluster without rewriting every workload, while still needing to redesign networking, identity and operational ownership for the cloud environment.

A useful inspection step is `kubectl get storageclass,pvc -A`. Read the output as evidence, not as a ritual: first confirm the desired object exists, then look at status conditions, events and the Google Cloud resource it represents. In production, capture the expected result in a runbook or automated check so an operator can distinguish slow reconciliation from a configuration error.

Production gotcha: Access mode is only one dimension; latency, zone scope, consistency and restore behavior also matter. The safe habit is to verify quotas, regional availability and feature support against current Google Cloud documentation before rollout. Start with block storage.