Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-33046

When we scale up new nodes, the PinnedImageSets status in the MCPs takes very long to be updated


      Description of problem:

      When new nodes are scaled up, the PinnedImageSets status in the MCP takes a long while to be updated.
      Nevertheless, after a long time (sometimes about 30mins) they are correctly updated.

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

      pre-merge: https://github.com/openshift/machine-config-operator/pull/4303

      How reproducible:


      Steps to Reproduce:

          1. Create a pinnedimageset for the worker node with one only image (so that it is fast)
          2. Wait until the pinned image is pinned in all worker nodes
          3. Scale up 10 new worker nodes, or thereabouts

      Actual results:

      When all new nodes are created, the image is correctly pinned in all of them, but the status in the MCP is not fully synced until a long time later. Sometimes even 25 or 30 minutes.
      oc get mcp worker -o yaml
        - availableMachineCount: 12
          machineCount: 12
          poolSynchronizerType: PinnedImageSets
          readyMachineCount: 12
          unavailableMachineCount: 0
          updatedMachineCount: 10
        readyMachineCount: 12

      Expected results:

      The status in the MCP should be updated earlier once all nodes have finished pinning the image.

      Additional info:


            sbatschelet Sam Batschelet
            sregidor@redhat.com Sergio Regidor de la Rosa
            Sergio Regidor de la Rosa Sergio Regidor de la Rosa
            0 Vote for this issue
            3 Start watching this issue