Sunny

/easycla

Is there any ongoing effort to work on this feature support?

Related pr in filestore for reference: https://github.com/kubernetes-sigs/gcp-filestore-csi-driver/pull/425

/ok-to-test

/sig storage /kind bug

Let's continue the conversation here. @msau42 @jsafrane If the provisioning has started, shouldn't the state changed to ProvisioningInBackground? If it is ProvisioningInBackground and node is not found, we do **not**...

@jsafrane we are trying to solve the case that the provisioning did start already on the backend but then the node got deleted afterwards.

The node with the same node name will not come back if it gets deleted. But if a node with the same node name but a different UID may come...

What happens if a node with the exact name but different uid comes back? Does this shorter string provide enough information for the provisioner to distinguish the node and decide...

But how do we know if a node will come back or not? How do we know if it will be recreated from symptoms? Let say we store the nodeName...