Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-52878

Kubernetes API Server apply-bootstrap container does not respect SIGTERM

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done-Errata
    • Icon: Normal Normal
    • None
    • 4.15, 4.16, 4.17, 4.18
    • HyperShift
    • None
    • Moderate
    • None
    • False
    • Hide

      None

      Show
      None
    • Hide
      * Previously, the `apply-bootstrap` container did not handle a `SIGTERM` signal correctly. The container waited for a sleep operation to complete before handling the signal, which then exceeded the `termination-grace-period` of the pod. This situation required a `SIGKILL` signal to force the shutdown operation and allow the pods to finish deletion. With this release, the `apply-bootstrap` container now handles the signal `SIGTERM` correctly so a correct graceful shutdown period happens without the need for a `SIGKILLED` signal. (link:https://issues.redhat.com/browse/OCPBUGS-52878[*OCPBUGS-52878*])
      Show
      * Previously, the `apply-bootstrap` container did not handle a `SIGTERM` signal correctly. The container waited for a sleep operation to complete before handling the signal, which then exceeded the `termination-grace-period` of the pod. This situation required a `SIGKILL` signal to force the shutdown operation and allow the pods to finish deletion. With this release, the `apply-bootstrap` container now handles the signal `SIGTERM` correctly so a correct graceful shutdown period happens without the need for a `SIGKILLED` signal. (link: https://issues.redhat.com/browse/OCPBUGS-52878 [* OCPBUGS-52878 *])
    • Bug Fix
    • Done

      Description of problem:

          The apply-bootstrap container does not terminate within the `TerminationGracePeriodSeconds` timer. This is because the container does not respect a SIGTERM signal. This is an issue to address this bug and ensure the container respects a SIGTERM when issued and thus terminates within the window.

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

      How reproducible:

      Steps to Reproduce:

          1. Create hypershift enabled cluster
    2. Delete API server pods and observe the apply-bootstrap container be force killed instead of terminating gracefully.
    3.

      Actual results:

      Expected results:

      Additional info:

              jared.hayes Jared Hayes (Inactive)
              jared.hayes Jared Hayes (Inactive)
              Wen Wang Wen Wang
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated:
                Resolved: