neko1437 commented on issue #47601: URL: https://github.com/apache/airflow/issues/47601#issuecomment-2952231722
### Apache Airflow Version - `2.11.0` ### Apache Airflow Providers Versions - `apache-airflow-providers-cncf-kubernetes==10.5.0` - `apache-airflow-providers-google==15.1.0` --- ### Issue Description I encountered the same issue when using the `GKEStartPodOperator` with `on_finish_action="delete_pod"`. Despite this configuration, the pod does **not** get deleted at the end of a task run. The task logs contain messages like: ``` "Skipping deleting pod ..." ``` --- ### Diagnostics To isolate the cause, I downgraded the `apache-airflow-providers-google` package step by step. I found that the issue was introduced in **version 12.0.0**. Everything works as expected when using: - `apache-airflow-providers-cncf-kubernetes==10.5.0` - `apache-airflow-providers-google==11.0.0` With this combination, the pod is correctly deleted at the end of the task. --- ### Impact This issue significantly affects task reliability. When a task fails and retries are enabled, the retry attempt gets **skipped** because it detects an existing pod (with the same labels) that was not cleaned up. This behavior undermines Airflow’s retry mechanism and can lead to missed or incomplete processing. Given the potential for skipped retries and unintended side effects, I believe this is a **critical issue** that should be addressed. -- 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]
