Uploaded image for project: 'Knative Serving'
  1. Knative Serving
  2. SRVKS-785

[net-istio] Connections closed with HTTP 503, "upstream connect error or disconnect/reset before headers. reset reason: connection termination" on graceful shutdown, ignores revision timeoutSeconds

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Undefined Undefined
    • 1.31.0
    • 1.17.0, 1.18.0, 1.21.0
    • None
    • None
    • False
    • False
    • Undefined

      With Serverless 1.16 / 1.17, with net-istio setup, having a ksvc with `timeoutSeconds: 45`

      apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: autoscale
  annotations:
    serving.knative.openshift.io/enablePassthrough: "true"
spec:
  template:
    metadata:
      annotations:
        sidecar.istio.io/inject: "true"
        sidecar.istio.io/rewriteAppHTTPProbers: "true"
    spec:
      timeoutSeconds: 45
      containers:
      - image: quay.io/openshift-knative/serving/autoscale:v1.10

      issuing requests like

      curl https://autoscale-foobar.apps.example.com/?sleep=12001

      and deleting a pod while a request is still ongoing

      Causes the request to be cut ungracefully with a HTTP 503 "upstream connect error or disconnect/reset before headers. reset reason: connection termination" error (despite the timeoutSeconds set)

        1. reproducer-pod-delete-503-on-long-request.sh
          1 kB

              Unassigned Unassigned
              maschmid@redhat.com Marek Schmidt
              Votes:
              0 Vote for this issue
              Watchers:
              7 Start watching this issue

                Created:
                Updated:
                Resolved: