Before deploymentkubernetes-storage-readiness

Kubernetes Storage Readiness Checklist

13 items ·9 critical ·3 warn ·1 info

How to use this checklist

Storage is where a Kubernetes mistake becomes permanent. Everything else can be redeployed from Git.

Mark an item N/A when it genuinely does not apply, and write down why. An unexplained N/A is the most common way a checklist stops working.

Sign-off

Every critical item must pass. A failing critical item blocks the deployment or the maintenance window; it is not a note for later. Record the date, the reviewer, and the disposition of every item that did not pass.

Critical9 items

The CSI driver controller and node plugins are Running, and the driver version is supported against this Kubernetes version.
```
kubectl get csidrivers && kubectl get pods -A -l app.kubernetes.io/name=<csi-driver>
```
Exactly one StorageClass is marked default. Zero leaves PVCs Pending forever; two is undefined behaviour.
```
kubectl get storageclass
```
Every StorageClass reclaim policy is deliberate and documented. `Delete` means deleting a PVC destroys the data.
```
kubectl get storageclass -o custom-columns=NAME:.metadata.name,RECLAIM:.reclaimPolicy
```
Access modes match the workload. ReadWriteOnce cannot be shared between Pods on different nodes, and the failure appears only at scale-out.
volumeBindingMode is WaitForFirstConsumer for zonal storage, so a volume is not provisioned in a zone the Pod cannot reach.
The snapshot CRDs and controller are installed, a VolumeSnapshotClass exists, and a snapshot has been taken and restored at least once.
```
kubectl get volumesnapshotclass
```
Volume snapshots are not the only backup. A snapshot that lives in the same storage system as the volume does not survive that system failing.
A full restore of a stateful workload has been rehearsed end to end within the last quarter, and the elapsed time was measured.
Alerts fire on volume fullness before the application notices, at a threshold that leaves time to act.

Warning3 items

No PersistentVolume still carries the deprecated `Recycle` reclaim policy. It is accepted by the API and implemented by nothing.

kubectl get pv -o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.spec.persistentVolumeReclaimPolicy}{"\n"}{end}' | grep Recycle

allowVolumeExpansion is enabled where the underlying storage supports it, and expansion has been tested once.
The behaviour of PVCs on StatefulSet deletion is understood and documented — they are retained by default, and they accumulate.
```
kubectl get pvc -A
```

Info1 item

Released PersistentVolumes are audited so that orphaned storage does not accumulate cost and confusion.
```
kubectl get pv --field-selector=status.phase=Released
```