PersistentVolume attachment issue with IBM Block Storage CSI Driver

Issue

• During a node drain operation, or a simple restart of a Pod to start on a different node, if the pod consumes a volume provisioned from the IBM Block Storage CSI driver, it could hang indefinitely in the detach operation.
• On SVC storage, the corresponding LUN stays mapped on the old worker node. Manually deleting the mapping on SVC results in an immediate remapping of the LUN to the old worker node again.