I am creating a deployment that is going to use a GCE persistent disk to create a volume for the deployment's pod to use.
Since the pod is ephemeral, it may so happen that upon restarts, it ends up on a different node than the PD/PV/PVC.
In that case the following error may occur
Multi-Attach error for volume Volume is already exclusively attached
I am aware of this answer.
My question is whether there is a way of predictably / consistently collocating the pod that needs the volume and the PD on the same node so as to avoid restart delays.