Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-44911

Cluster installation is deemed complete when all control plane nodes have not joined

XMLWordPrintable

    • Important
    • None
    • Proposed
    • False
    • Hide

      None

      Show
      None

      Description of problem:

          The wait-for install-complete command deems a cluster has been successful installed by checking that all cluster operators are available. For a SNO or HighlyAvailable cluster, this is achieved when the operator's pod replica is 1 or 3.
          Because the same HighlyAvailable cluster topology is used with control plane replicas set to 4 or 5, it is possible for the wait-for install-complete command to report the cluster installation is successful when all control plane nodes have not joined or are not in ready status. This is because the cluster operators are available when it sees a replica count of 3. One would expect a cluster operator to be available when its pod replica count matches the control plane replica count (4 or 5).  

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

          4.18.0

      How reproducible:

          Always

      Steps to Reproduce:

          1. Install a 4 or 5 control plane replica cluster, but only bring up 4-1 or 5-1 nodes.
          2. Run wait-for install-complete
          3.
          

      Actual results:

          Installation is reported to be successful with 3 (4 replicas desired) or 5 nodes (5 replicas desired).

      Expected results:

          Installation is reported successful when the number of control plane nodes matches the replica count specified in the install-config.

      Additional info:

          

              rwsu1@redhat.com Richard Su
              rwsu1@redhat.com Richard Su
              Biagio Manzari Biagio Manzari
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated: