[Azure] Customer upgraded to 4.13.34 from 4.13.30, RBAC issues for system:nodes unable to create serviceaccount/token

    Cluster is unable to start up openshift-image-registry pods and openshift-logging pods with the error message

142m        Warning   FailedMount              pod/image-registry-689549b84f-dk885                     MountVolume.SetUp failed for volume "bound-sa-token" : failed to fetch token: serviceaccounts "registry" is forbidden: User "system:node:worker-node-7qh98" cannot create resource "serviceaccounts/token" in API group "" in the namespace "openshift-image-registry": no relationship found between node 'worker-node-7qh98' and this 

object142m        Warning   FailedMount              pod/image-registry-689549b84f-dk885                     MountVolume.SetUp failed for volume "kube-api-access-8ksnc" : failed to fetch token: serviceaccounts "registry" is forbidden: User "system:node:worker-node-7qh98" cannot create resource "serviceaccounts/token" in API group "" in the namespace "openshift-image-registry": no relationship found between node 'worker-node-7qh98' and this object

    4.13.35

    Unknown, first customer to see this issue

Customers' upgrade path in finding this bug:    

    1. Install cluster at 4.12.44
    2. Upgrade to 4.13.30
    3. Upgrade to 4.13.34

Actual results:
{code:none}
    Image registry pods and logging pods continously recreate every 5-10 seconds never becoming healthy

    Image registry and logging pods are able to create and run normally

    This is the first time SRE is seeing this issue for a customer. Unforunately SRE has not identified a workaround we have:

1. Deleted and recreated the image-registry deployment several times
2. Updated the node selection for image-registry pods to be all nodes. This issue perists on all nodes on the cluster, even master nodes
3. Inspected the role bindings of the cluster, system:node has the correct permissions