Uploaded image for project: 'OpenShift Runtimes'
  1. OpenShift Runtimes
  2. RUN-1776

kube play: service container: exit code propagation

XMLWordPrintable

    • Icon: Story Story
    • Resolution: Done
    • Icon: Undefined Undefined
    • None
    • None
    • None
    • 3
    • False
    • None
    • False
    • sst_container_tools
    • RUN 235, RUN 236, RUN 237, RUN 238

      When running `podman kube play` in systemd, the so-called "service container" is used to track the lifecycle of the service in systemd and acts as the service's  main PID (see blog [1]).

      The service container will exit once all containers (and hence pods) in the service have exited.  However, the service container will always exit with zero as there is no link yet between the exit codes of the containers and the service container.

      The goal of this issue is to implement a mechanism (opt-in) to let the service container container exit non-zero if containers inside the service have failed/exited non-zero. 

      Please refer to the design document for more details: https://docs.google.com/document/d/1waYuiMwAQckwHB9dVq0DIYLtvMZe2sl0lxRoz8UXHLQ/edit?usp=sharing

            vrothber@redhat.com Valentin Rothberg
            vrothber@redhat.com Valentin Rothberg
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: