Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-5821

system:controller:ttl-controller referencing non existing serviceAccount

    XMLWordPrintable

Details

    • Low
    • False
    • Hide

      None

      Show
      None

    Description

      Description of problem:

      $ oc get clusterversion
      NAME      VERSION   AVAILABLE   PROGRESSING   SINCE   STATUS
      version   4.11.20   True        False         43h     Cluster version is 4.11.20
      
      $ oc get clusterrolebinding system:controller:ttl-controller -o yaml
      apiVersion: rbac.authorization.k8s.io/v1
      kind: ClusterRoleBinding
      metadata:
        annotations:
          rbac.authorization.kubernetes.io/autoupdate: "true"
        creationTimestamp: "2023-01-11T13:09:19Z"
        labels:
          kubernetes.io/bootstrapping: rbac-defaults
        name: system:controller:ttl-controller
        resourceVersion: "204"
        uid: 74536183-0c18-44eb-9ce2-afcd85542580
      roleRef:
        apiGroup: rbac.authorization.k8s.io
        kind: ClusterRole
        name: system:controller:ttl-controller
      subjects:
      - kind: ServiceAccount
        name: ttl-controller
        namespace: kube-system
      
      $ oc get sa ttl-controller -n kube-system
      Error from server (NotFound): serviceaccounts "ttl-controller" not found
      
      The serviceAccount ttl-controller does not exist. Neither in kube-system nor in any other namespace.
      
      It's therefore not clear what this ClusterRoleBinding does, what use-case it does fulfill and why it references non existing serviceAccount.
      
      From Security point of view, it's recommended to remove non serviceAccounts from ClusterRoleBindings as a potential attacker could abuse the current state by creating the necessary serviceAccount and gain undesired permissions.

      Version-Release number of selected component (if applicable):

      OpenShift Container Platform 4 (all version from what we have found)

      How reproducible:

      Always

      Steps to Reproduce:

      1. Install OpenShift Container Platform 4
      2. Run oc get clusterrolebinding system:controller:ttl-controller -o yaml

      Actual results:

      $ oc get clusterrolebinding system:controller:ttl-controller -o yaml
      apiVersion: rbac.authorization.k8s.io/v1
      kind: ClusterRoleBinding
      metadata:
        annotations:
          rbac.authorization.kubernetes.io/autoupdate: "true"
        creationTimestamp: "2023-01-11T13:09:19Z"
        labels:
          kubernetes.io/bootstrapping: rbac-defaults
        name: system:controller:ttl-controller
        resourceVersion: "204"
        uid: 74536183-0c18-44eb-9ce2-afcd85542580
      roleRef:
        apiGroup: rbac.authorization.k8s.io
        kind: ClusterRole
        name: system:controller:ttl-controller
      subjects:
      - kind: ServiceAccount
        name: ttl-controller
        namespace: kube-system
      
      $ oc get sa ttl-controller -n kube-system
      Error from server (NotFound): serviceaccounts "ttl-controller" not found

      Expected results:

      The serviceAccount called ttl-controller to exist or otherwise the ClusterRoleBinding to be removed.

      Additional info:

      Finding related to a Security review done on the OpenShift Container Platform 4 - Platform

      Attachments

        Issue Links

          Activity

            People

              fkrepins@redhat.com Filip Krepinsky
              rhn-support-sreber Simon Reber
              Zhaohua Sun Zhaohua Sun
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: