kubernetes: add configure option to disable shell detection entrypoint
What does this MR do?
Allows user to configure runner pods to preserve default container image entrypoints instead of the runner overriding them with shell detection logic.
Why was this MR needed?
Kubernetes executor behavior differs from docker executor behavior by not preserving default image entrypoints. This behavior maybe advantageous for some images that do not typically allow pods to accept shell input from the CI script. However, a container image could be configured to provide a shell environment for the CI script to successfully execute, and the container image may perform other initialization in its entrypoint that it wishes to maintain.
Are there points in the code the reviewer needs to double check?
Does this MR meet the acceptance criteria?
-
Documentation created/updated -
Added tests for this feature/bug -
In case of conflicts with master
- branch was rebased