ferruzzi commented on PR #71732: URL: https://github.com/apache/airflow/pull/71732#issuecomment-5592749827
@arose26 - Please hold on this one for a moment, and apologies for the three weeks of silence; that's on us, not on you. Your diagnosis here is the right one and I want to keep it. #66350 fixes the same crash, but for missed part that you caught. They stop the crash by removing the filter rather than fixing it, so `airflow db clean --dag-ids one_dag` would quietly clean those two tables for every dag. Your `dag_id_via` approach filters through the parent's `dag_id` instead, which is what should ship. Your `EXISTS` over `NOT IN` is right too, since `deadline.dagrun_id` is nullable. I also appreciate that you checked the rest of `config_list` against the table metadata instead of trusting the issue's list to be complete. The problem is that the two PRs are each half of a fix, and #66350 already carries the `Airflow 3.3.2` milestone while this one carries none. If they land separately, or if one gets backported without the other, we ship a release where `--dag-ids` silently over-deletes. So I've asked over on #66350 whether they can cherry-pick your changes in with a co-author trailer and land the whole thing as one PR. I'm adding the `Airflow 3.3.2` milestone here in the meantime, purely so this can't drift out of the patch release while we decide. -- 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]
