Describe the issue:
In our product docs, we have some bits[1] and pieces[2] of what is
required to run OCS on tainted OCP nodes, but the information is not complete and spread out over multiple locations. 

[1]
https://access.redhat.com/documentation/en-us/red_hat_openshift_container_storage/4.6/html/managing_and_allocating_storage_resources/how-to-use-dedicated-worker-nodes-for-openshift-container-storage_rhocs#adding_toleration_for_local_storage_operator

[2]
https://access.redhat.com/documentation/en-us/red_hat_openshift_container_storage/4.6/html/managing_and_allocating_storage_resources/managing-container-storage-interface-component-placements_rhocs


Suggestions for improvement:
First and foremost, we should gather it all in one place and not spread the information out over multiple chapters.

Second, the current information focuses on "infra" nodes. And while
infra-nodes is a special-case, the same applies to any kind of taints which are applied.
So we should make this clear that this procedure applies as soon as one want to have a taint on your nodes for any kind of reason. And one such reason might be that you want to define an infra node.

Third, we're entirely missing that toleatrions need to be added to LocalVolumeDiscovery CR for this to work with auto-discovered volumes, see example below:

apiVersion: local.storage.openshift.io/v1alpha1
kind: LocalVolumeDiscovery
metadata:
  name: auto-discover-devices
  namespace: openshift-local-storage
spec:
  tolerations:
  - effect: NoSchedule
    key: myTaint
    value: myValue