Uploaded image for project: 'Apache YuniKorn'
  1. Apache YuniKorn
  2. YUNIKORN-1153

Admission controller: first health check should be delayed

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Open
    • Minor
    • Resolution: Unresolved
    • None
    • None
    • shim - kubernetes
    • None

    Description

      When deploying Yunikorn locally, I often see the first health check failing:

      Events:
        Type     Reason     Age                   From               Message
        ----     ------     ----                  ----               -------
        Normal   Scheduled  3m12s                 default-scheduler  Successfully assigned default/yunikorn-admission-controller-78c775cfd9-6pp8d to minikube
        Normal   Pulled     3m12s                 kubelet            Container image "apache/yunikorn:admission-latest" already present on machine
        Normal   Created    3m12s                 kubelet            Created container yunikorn-admission-controller
        Normal   Started    3m11s                 kubelet            Started container yunikorn-admission-controller
        Warning  Unhealthy  2m52s (x2 over 3m2s)  kubelet            Startup probe failed: Get "https://192.168.49.2:9089/health": dial tcp 192.168.49.2:9089: connect: connection refused
      

      We need to add some initialDelaySeconds to wait with the first probe. 10-15 seconds is probably a good value.

      Attachments

        Activity

          People

            lowc1012 Ryan Lo
            pbacsko Peter Bacsko
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated: