PersistentVolumeClaim stuck in Pending
A Pending PVC has no matching volume. Check the claim's events, its storage class and access mode, and whether a default StorageClass exists.
What you see
kubectl get pvcshowsPending- The pod that mounts the claim stays
Pending
What it means
A claim is a request for storage with a size, an access mode and, optionally, a storage class. The control plane binds it to a volume that fits. The docs state that claims remain unbound indefinitely if a matching volume does not exist.
With a StorageClass and a provisioner, the volume is created on demand. Without one, an administrator must have created a matching volume in advance.
How to confirm the cause
- 1Read the claim's events
kubectl describe pvc CLAIM_NAME
The Events section usually says why it is not bound.
- 2See available volumes and their classes
kubectl get pv && kubectl get storageclass
Check for a volume that matches the size, access mode and class, and for a default StorageClass (marked
(default)). - 3See which pod is waiting on it
kubectl get pods -o wide
The pod stays Pending until the claim binds.
Causes and fixes
No volume matches and nothing provisions one
- HOW TO TELL
- No PV fits and there is no StorageClass with a provisioner.
- ROOT CAUSE
- Static binding needs a pre-created PV with enough size, the same access modes and the same class. Without a provisioner, none is created automatically.
- FIX
- Create a matching PV, or install a StorageClass with a provisioner and reference it in the claim.
The storage class name does not match
- HOW TO TELL
- The claim names a class that does not exist, or sets
storageClassName: "". - ROOT CAUSE
- A claim with no class binds only to a PV of the default class, and one with an empty string binds only to PVs with no class. Neither matches a class that is not there.
- FIX
- Set
storageClassNameto a class that exists, or fix the default class.
Access mode cannot be satisfied
- HOW TO TELL
- The claim asks for
ReadWriteManyand the storage only supportsReadWriteOnce. - ROOT CAUSE
ReadWriteOncelets the volume be mounted read-write by a single node. A claim that needs access from several nodes cannot use it.- FIX
- Use storage that supports the access mode you need, or change the claim.
How to stop it coming back
- Know your cluster's default StorageClass and name it explicitly in manifests.
- Match access modes to what the storage backend supports.
- Alert on claims that stay Pending.
Keep going
- Runbook: Pod stuck in Pending with FailedScheduling
- Cheat sheet: kubectl
- Lesson: the kubernetes topic in Learn
Sources
Kubernetes documentation (kubernetes.io), read 18 Nov 2026. The docs version selector listed v1.33 to v1.37.