FailedMount
The kubelet couldn’t mount one of the pod’s volumes, so the pod stays in ContainerCreating. The message usually arrives after a two-minute timeout and names the volume.
Diagnose in order
Section titled “Diagnose in order”- Which volume, and which step?
MountVolume.SetUpfailures forconfigmaporsecretvolumes mean a missing object.AttachVolumefailures are the attach step.MountDeviceorNodeStagefailures are on the node. - For CSI volumes, check the attachment:
kubectl get volumeattachment | grep <pv-name>.ATTACHED falsepoints at the CSI controller or the storage backend. - Check the CSI node plugin pod on that node and its logs.
Usual causes
Section titled “Usual causes”| Message | Cause |
|---|---|
configmap "x" not found / secret "x" not found |
The object referenced as a volume doesn’t exist in the namespace |
Waiting for detach for volume (older: Multi-Attach error) |
The RWO disk is still attached to another node; see Multi-Attach error |
AttachVolume.Attach failed |
Backend refused or timed out: disk limit per node, wrong zone, driver credentials |
MountDevice failed ... DeadlineExceeded |
Node plugin hung or the device never appeared on the node |
NFS: mount.nfs: access denied / Connection timed out |
Export permissions or network path to the NFS server |
Create the missing object, or fix the storage path the message names. If a single node keeps failing to mount, cordon it and restart its CSI node plugin pod.
Read next: Storage and CSI.