Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-74093

The WICD controller should periodically reconcile the state of Windows services and automatically restart any that have stopped. This is failing on GCP clusters

    • Icon: Bug Bug
    • Resolution: Cannot Reproduce
    • Icon: Major Major
    • None
    • 4.21.0
    • Windows Containers
    • None
    • None
    • False
    • Hide

      None

      Show
      None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Description of problem:

          The WICD controller should periodically reconcile the state of Windows
        services and automatically restart any that have stopped. This is failing
        on GCP clusters.

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

          4.21-83ede63

      How reproducible:

      Most likely    

      Steps to Reproduce:

        1. Deploy cluster with Windows nodes on GCP/AWS
        2. Manually stop windows_exporter service: sc.exe stop windows_exporter
        3. Wait 10 minutes
        4. Check service state: sc.exe query windows_exporter
           

      Actual results:

          Service remains stopped after 10 minutes

      Expected results:

        WICD controller should:
         1. Periodically check service states on Windows nodes
         2. Compare actual state vs desired state (from ConfigMap)
         3. Restart any services that are stopped
         4. Log reconciliation actions   

      Additional info:

       =================================================================
         WICD CONFIGMAP
         =================================================================
         ConfigMap: windows-services-10.21.0-83ede63
         Services Defined:
         1. containerd (priority: 0, bootstrap)
         2. kubelet (priority: 1, bootstrap)
         3. windows_exporter (priority: 2) ← TEST TARGET
         4. hybrid-overlay-node (priority: 2)
         5. csi-proxy (priority: 2)
         6. kube-proxy (priority: 3) 
      ConfigMap is correctly defined but no reconciliation loop is monitoring service states.
      
       

              rh-ee-ssoto Sebastian Soto
              rrasouli Aharon Rasouli
              None
              None
              Aharon Rasouli Aharon Rasouli
              None
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                Resolved: