Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-30986

Hypershift hosted cluster does not work with old tlsSecurityProfile

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Normal Normal
    • None
    • 4.14.z, 4.15.z, 4.16.0
    • HyperShift
    • Critical
    • Yes
    • Hypershift Sprint 251, Hypershift Sprint 252, Hypershift Sprint 253, Hypershift Sprint 254, Hypershift Sprint 255, Hypershift Sprint 256
    • 6
    • Rejected
    • False
    • Hide

      None

      Show
      None

      Description of problem:

      After we applied the old tlsSecurityProfile to the Hypershift hosted clsuter, the apiserver ran into CrashLoopBackOff failure, this blocked our test.
          

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

      $ oc get clusterversion
      NAME      VERSION                              AVAILABLE   PROGRESSING   SINCE   STATUS
      version   4.16.0-0.nightly-2024-03-13-061822   True        False         129m    Cluster version is 4.16.0-0.nightly-2024-03-13-061822
      
          

      How reproducible:

          always
          

      Steps to Reproduce:

          1. Specify KUBECONFIG with kubeconfig of the Hypershift management cluster
          2. hostedcluster=$( oc get -n clusters hostedclusters -o json | jq -r .items[].metadata.name)
          3. oc patch hostedcluster $hostedcluster -n clusters --type=merge -p '{"spec": {"configuration": {"apiServer": {"tlsSecurityProfile":{"old":{},"type":"Old"}}}}}'
      hostedcluster.hypershift.openshift.io/hypershift-ci-270930 patched
          4. Checked the tlsSecurityProfile,
          $ oc get HostedCluster $hostedcluster -n clusters -ojson | jq .spec.configuration.apiServer
      {
        "audit": {
          "profile": "Default"
        },
        "tlsSecurityProfile": {
          "old": {},
          "type": "Old"
        }
      }
          

      Actual results:

      One of the kube-apiserver of Hosted cluster ran into CrashLoopBackOff, stuck in this status, unable to complete the old tlsSecurityProfile configuration.
      
      $ oc get pods -l app=kube-apiserver  -n clusters-${hostedcluster}
      NAME                              READY   STATUS             RESTARTS      AGE
      kube-apiserver-5b6fc94b64-c575p   5/5     Running            0             70m
      kube-apiserver-5b6fc94b64-tvwtl   5/5     Running            0             70m
      kube-apiserver-84c7c8dd9d-pnvvk   4/5     CrashLoopBackOff   6 (20s ago)   7m38s
          

      Expected results:

          Applying the old tlsSecurityProfile should be successful.
          

      Additional info:

         This also can be reproduced on 4.14, 4.15. We have the last passed log of the test case as below:
        passed      API_Server       2024-02-19 13:34:25(UTC)    aws 	4.14.0-0.nightly-2024-02-18-123855   hypershift 	
        passed      API_Server	  2024-02-08 02:24:15(UTC)   aws 	4.15.0-0.nightly-2024-02-07-062935 	hypershift
        passed      API_Server	  2024-02-17 08:33:37(UTC)   aws 	4.16.0-0.nightly-2024-02-08-073857 	hypershift
      
      From the history of the test, it seems that some code changes were introduced in February that caused the bug.
          

            sjenning Seth Jennings
            wk2019 Ke Wang
            Ke Wang Ke Wang
            Votes:
            0 Vote for this issue
            Watchers:
            7 Start watching this issue

              Created:
              Updated: