I am experiencing issues with PVC’s not detaching from nodes in my EKS cluster. This occurs sporadically, but it’s most notable during helm upgrades. For example, when I upgrade Harbor roughly half of the pods are stuck in an “Init:0/1” state with an error like:

Warning FailedAttachVolume 9m55s attachdetach-controller Multi-Attach error for volume "pvc-agv3sv3sa-fsdr-gsd4-be20-avav3a35dsf3" Volume is already exclusively attached to one node and can't be attached to another

I have tried setting terminationGracePeriodSeconds to 0 for the deployments, and modified the updateStrategy from Rolling to Recreate. Unfortunately, these do not help with the PVC’s. The typical solution I have is to manually detach the volume in AWS, which is not ideal.