Abdulrehman-PIAIC80387 commented on issue #68699: URL: https://github.com/apache/airflow/issues/68699#issuecomment-4738816856
Root cause looks like an atomicity issue rather than anything SQLite-specific: `_create_backfill` commits the `Backfill` row ([backfill.py#L678](https://github.com/apache/airflow/blob/main/airflow-core/src/airflow/models/backfill.py#L678)) **before** creating its runs. If run creation then fails (the `database is locked` here, but any error would do), the committed row survives with no/partial runs, and the `num_active > 0` check at L652 blocks all future backfills for that dag. Two ways to fix it — which would you prefer? 1. **Atomic creation** (`flush` instead of the early `commit` so it all rolls back together) **+ a real "one active backfill per dag" guard** (row lock / unique index) — since that early commit currently doubles as a concurrency guard. 2. **Cleanup-on-failure** — keep the commit, but delete the orphaned `Backfill` row if run creation fails. Happy to open a PR for whichever direction you prefer. -- 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]
