Taragolis commented on issue #36809:
URL: https://github.com/apache/airflow/issues/36809#issuecomment-1895499822
I check on the `main` and this kind of bug exists.
However it not only affects `ShortCircuitOperator`, I reproduced on regular
task flow operators
```python
from __future__ import annotations
from typing import TYPE_CHECKING
from airflow import DAG
from airflow.decorators import task
from airflow.utils.task_group import TaskGroup
if TYPE_CHECKING:
from airflow.models import TaskInstance
with DAG(dag_id="issue_36809", schedule=None, tags=["issue", "36809",
"zombie?"]):
@task(retries=1, retry_delay=5)
def flakey_failure(ti: TaskInstance = None) -> int:
if (try_number := ti.try_number) == 1:
raise SystemExit(-1)
return try_number
@task
def my_task(i: int):
print(f"Upstream try number: {i}")
for ix in range(1, 10 + 1):
with TaskGroup(group_id=f"group_{ix:02d}"):
my_task(flakey_failure())
```
Seems like the problem with zombie detection/resolution. In addition found
interesting bug, on next DagRun when task failed it also might marks (with high
probability) as failed even if it previously run successfully

--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]