In many respects, this issue we just described is very similar to the labels and selectors bug, but exhibits itself quite differently. The underlying system does its best to look for volumes, ConfigMaps, secrets, and so on, and to let you create them in any order. If you make a typo, or if a ConfigMap, secret, or volume isn't referenced correctly or is just missing, then the pod will ultimately fail.
These resources are all referenced dynamically. In doing that referencing, Kubernetes provides retries and timeouts, but cannot explicitly validate the failures before actually looking for the related resources and ultimately failing. This can make debugging these issues more time consuming. When you first look for why issues ...