Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-15812

Pod Garbage collector fails to clean up PODs from nodes that are not running anymore

XMLWordPrintable

    • No
    • False
    • Hide

      None

      Show
      None
    • Customer Facing

      Description of problem:

      A pod which has an environment variable specified twice is not cleaned up via garbage collection.

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

      4.13.2

      How reproducible:

      100%

      Steps to Reproduce:

      From https://access.redhat.com/support/cases/#/case/03552916/discussion?commentId=a0a6R00000V7VXfQAN
      
      1- create a pod that defines twice the same environment variable
      ```
      apiVersion: v1
      kind: Pod
      metadata:
        labels:
          test: liveness
        name: liveness-exec
      spec:
        nodeName: prd-ne-shop03-84q98-ssp-arm64-northeurope3-8bpm5
        containers:
        - name: liveness
          env:
          - name: toto
            value: tata
          - name: toto
            value: tata
          image: centos
          args:
          - /bin/sh
          - -c
          - sleep 7200
        tolerations:
        - effect: NoSchedule
          key: cpu
          value: arm64
        - effect: NoSchedule
          key: reservedFor
          value: ssp-arm64
      ```
      
      2- delete the node and machine where the pod runs
      3- verify that the pod still shows as ready (Kubernetes bug)

      Actual results:

      Pod still shows as ready

      Expected results:

      Pod is removed

      Additional info:

      https://github.com/kubernetes/kubernetes/issues/118261

            Unassigned Unassigned
            rhn-support-sdodson Scott Dodson
            Sunil Choudhary Sunil Choudhary
            Votes:
            0 Vote for this issue
            Watchers:
            8 Start watching this issue

              Created:
              Updated:
              Resolved: