A persistent volume (PV) is a cluster-scoped Kubernetes object for storage that outlives an individual pod. Keep data in a PV when replacing or deleting a pod must not erase it. The storage can be backed by a host or an external provider; the PV remains separate from the workload that mounts it. section 7.1
To make that storage available to an application:
- Create a PV with the required capacity, storage backing, volume mode, access mode, and reclaim policy. Filesystem is the default volume mode; Block is the alternative. section 7.1 section 7.5
- Create a persistent volume claim (PVC) to reserve storage, then mount the claim in a pod or Deployment. A claim can bind to a larger PV rather than requiring an exact capacity match. section 7.1 section 7.2
- Alternatively, choose a storage class in the PVC to provision a PV dynamically. Inspect the class with `kubectl get sc standard -o yaml`; automatically provisioned PVs inherit settings such as its reclaim policy and volume binding mode. section 7.3.1 section 7.5
Host-backed storage task
Create a 100Mi hostPath PV from `/mnt/data` on worker node `node01`. Claim 90Mi, then mount the claim in a Deployment at `/usr/share/nginx/html`. Check `kubectl get pv` and `kubectl get pvc`: the claim can show `Bound` to the 100Mi PV. Check `kubectl get deploy`, `kubectl get po`, and `kubectl describe po` to verify the workload and mount. section 7.1
root@kind-control-plane:/# kubectl get pvc NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE claim-02833 Bound vol02833 100Mi RWO manual 4m root@kind-control-plane:/# kubectl get pv NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS vol02833 100Mi RWO Retain Bound default/claim-02833 manual
The command output shows a 100Mi PV bound to its claim, including access mode, reclaim policy, and storage class. section 7.1
Watch the PV phase while troubleshooting: an Available PV can be claimed; a Bound PV is already reserved. A claim protecting storage used by a pod can leave attempted PV or PVC deletions in Terminating until the workload is removed. Access modes govern which nodes can use the volume and with what read/write permissions; the reclaim policy governs what happens when the storage is released. Choose these settings from the workload’s requirements, not just its requested capacity. section 7.2 section 7.5
