Overview:
When a secondary container fails, and no other steps fail, the entire container will not return a status of "failed". In this case, the parallel run will return a status of "canceled" and the job will have a status of "failed".
In the case that another step fails within the container, the entire container will return a status of "failed".
You should also be sure to check the indices of each executor, and ensure that you are able to see every executor's output.
Additional Resources:
Comments
Article is closed for comments.