Uploaded image for project: 'Red Hat OpenShift Data Science'
  1. Red Hat OpenShift Data Science
  2. RHODS-7921

Containers defined in the wrong order in the DSPipeline deployments

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Minor Minor
    • None
    • RHODS_1.25.0_GA
    • Pipelines
    • 1
    • False
    • Hide

      None

      Show
      None
    • False
    • None
    • Testable
    • No
    • No
    • No
    • Pending
    • None
    • ML Ops Sprint 1.29

      Description of problem:

      When I deploy a DataSciencePipelinesApplication named Sample, I see a deployment named ds-pipeline-sample. This deployment has two containers: oauth-proxy and ds-pipeline-api-server.

      Prerequisites (if any, like setup, operators/versions):

      Steps to Reproduce

      1. deploy a DataSciencePipelinesApplication application
      2. look at the deployment it creates, and the containers inside it

      Actual results:

      the oauth-proxy container appears first, then the ds-pipeline-api-server container

      Expected results:

      the main container of the pod should appear first, then the side-cars.
      This simplifies the inspection of the containers with oc logs, oc rsh, etc, which use the first container by default

      Reproducibility (Always/Intermittent/Only Once):

      Build Details:

      RHODS 1.25.0 RC

      Workaround:

      Additional info:

      I didn't look at the other deployments, they may suffer from the same problem.

       

      Disclaimer: ensure that the order of containers are not being referenced in any code (i.e. code that assumes oauth-proxy is the first container in the deployment spec. and also confirm nothing breaks when order is reversed)

              gfrasca-1 Giulio Frasca
              kpouget2 Kevin Pouget
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Created:
                Updated:
                Resolved: