Optimize subclasses of DummyOperator be set straight to success instead of being run #11393
Description
Apache Airflow version:
v2.0.0
What happened:
Custom operators inheriting from DummyOperator are go in to a scheduled state, where as DummyOperator itself is set straight to success.
What you expected to happen:
All operators should be successfully executed.
How to reproduce it:
from datetime import datetime
from airflow import DAG
from airflow.operators.dummy_operator import DummyOperator
class MyJob(DummyOperator):
template_fields_renderers = {
"body": "json"
}
template_fields = ("body",)
def __init__(self, body, *args, **kwargs):
super().__init__(*args, **kwargs)
self.body = body
with DAG(
dag_id="test",
schedule_interval=None,
start_date=datetime(2020, 8, 13),
) as dag:
MyJob(task_id="aaaa", body={"aaa": "bbb"})Anything else we need to know:
This is a known problem that we will fix, see: #10956 (comment)