Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-55898

Surface pod scheduling failures to the user instead of waiting forever

    XMLWordPrintable

    Details

    • Similar Issues:

      Description

      Every now and then developers will either make environmental changes or changes to the pod spec that cause the pod to fail to schedule. I've seen this most frequently happen when:

      • A developer removes a container image from an internal registry that is referenced in the pod spec without updating the pod spec in all branches (The container will get stuck in ErrImagePull)
      • A developer adds a container to the pod spec without overriding the entrypoint. When the container exits, kubernetes kills and restarts the pod before jenkins sees the pod becoming ready

      It would be awesome if the kubernetes plugins could detect these failures and fail the build if this happens so the build doesn't run forever and we don't block other builds from scheduling. This is especially useful for users who don't have access to the kubernetes api to inspect *why* their builds aren't running.

        Attachments

          Activity

          Hide
          csanchez Carlos Sanchez added a comment -

          There are some screens in the node executor page but it can't be printed in the build logs because agent provisioning is orthogonal to the build

          Show
          csanchez Carlos Sanchez added a comment - There are some screens in the node executor page but it can't be printed in the build logs because agent provisioning is orthogonal to the build

            People

            • Assignee:
              csanchez Carlos Sanchez
              Reporter:
              nlowe_hyland Nathan Lowe
            • Votes:
              1 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated: