Unable to mount an Hetzner Volume

Steps

Diagnostic

Check the events

Kubernetes records most provisioning and mounting problems as Events:

• on the namespace
• on the PersistentVolumeClaim or the Pod:

Look at the Events section at the bottom of the output.

kubee -n namespace kubectl describe pvc <PVC-NAME>
kubee -n namespace kubectl describe pod <POD-NAME>

Check the controller logs

The controller is composed of the following container:

• Csi Driver (main container)

kubectl logs \
  -n kube-system -l app.kubernetes.io/name=hcloud-csi,app.kubernetes.io/component=controller \
  -c hcloud-csi-driver

• Resizer (sidecar)

kubectl logs \
  -n kube-system -l app.kubernetes.io/name=hcloud-csi,app.kubernetes.io/component=controller \
  -c csi-resizer

• Attacher (sidecar)

kubectl logs \
  -n kube-system -l app.kubernetes.io/name=hcloud-csi,app.kubernetes.io/component=controller \
  -c csi-attacher

https://github.com/hetznercloud/csi-driver/issues/46 https://github.com/k3s-io/k3s/issues/732#issuecomment-533896909

MountVolume.WaitForAttach failed for volume "combo-postgres-pv" :
volume 106135968 has GET error for
volume attachment csi-8eec466de784837604919617e7b87cd7746d2a2e1c0d712c9d1d06025825011e:
volumeattachments.storage.k8s.io "csi-8eec466de784837604919617e7b87cd7746d2a2e1c0d712c9d1d06025825011e"
is forbidden:
User "system:node:kube-server-01.eraldy.com" cannot get resource "volumeattachments"
in API group "storage.k8s.io" at the cluster scope: no relationship found between node 'kube-server-01.eraldy.com' and this object

2026-06-24T20:09:46.020271099Z E0624 20:09:46.018997       1 reflector.go:204] "Failed to watch" err="pods is forbidden: User \"system:serviceaccount:kube-system:hetzner-csi-controller\" cannot watch resource \"pods\" in API group \"\" at the cluster scope" logger="UnhandledError" reflector="k8s.io/client-go/informers/factory.go:161" type="*v1.Pod"

More Info

Troubleshooting guide

Known Error

The file /dev/disk/by-id/scsi-0HC_Volume_xxx does not exist

Check that the volume is not attached in the Hetzner console.

If not, the volumeattachment must be deleted if one exists. as proposed on this comment

Steps:

• Check that there is an attachment and get the name (ie csi-id) of the volumeattachment

kubee kubectl get volumeattachment | grep your-pvc-name

• Delete it if it exists

kubee kubectl delete volumeattachment csi-id
kubee kubectl delete volumeattachment csi-0f1e84ed8e2b316a4b746d0f0xxxx

Example output:

Warning: deleting cluster-scoped resources, not scoped to the provided namespace
volumeattachment.storage.k8s.io "csi-0f1e84ed8e2b316a4b746d0f0xxxx" deleted

• The volume should then appear in the list of disk.

# on the node
ls -la /dev/disk/by-id/

scsi-0HC_Volume_xxxx -> ../../sdb
scsi-0HC_Volume_xxxx -> ../../sdc