thc1006 opened a new pull request, #71749:
URL: https://github.com/apache/airflow/pull/71749
## What is wrong
The durable execution section of the KubernetesPodOperator doc say:
> The pod identity persisted in task state store isn't deleted
automatically, that only happens
> when someone runs ``airflow state-store clean``.
But `[state_store] clear_on_success` delete it too. When a task instance
reach SUCCESS, the execution API clear the whole task scope from the state
store:
```python
if updated_state == TaskInstanceState.SUCCESS:
if conf.getboolean("state_store", "clear_on_success"):
scope = TaskScope(...)
get_state_backend().clear(scope, session=session)
```
The default is `False`, so most people never see it. But someone who turn it
on read this sentence and think the pod identity still stay, and then the next
retry go to the label search instead of the direct reconnect, and they do not
know why.
## What I change
Only that one sentence. The rest of the paragraph, the part about
`retry_delay` and `default_retention_days`, I do not touch.
I find this while working on #71743, sorry for the small PR. Thank you for
reading it.
---
##### Was generative AI tooling used to co-author this PR?
- [x] Yes - Claude Code
This PR was written in part with the assistance of generative AI. I have
reviewed and tested every change myself.
--
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]
