robinedwards opened a new issue #18801: URL: https://github.com/apache/airflow/issues/18801
### Apache Airflow version 2.2.0b2 (beta snapshot) ### Operating System Debian GNU/Linux 11 (bullseye) ### Versions of Apache Airflow Providers ``` apache-airflow-providers-amazon @ file:///root/.cache/pypoetry/artifacts/c9/69/16/ffa2eb7a2e6e850a7048eaf66b6c40c990ef7c58149f20d3d3f333a2e9/apache_airflow_providers_amazon-2.2.0-py3-none-any.whl apache-airflow-providers-celery @ file:///root/.cache/pypoetry/artifacts/6e/1b/2f/f968318a7474e979af4dc53893ecafe8cd11a98a94077a9c3c27304eb7/apache_airflow_providers_celery-2.1.0-py3-none-any.whl apache-airflow-providers-ftp @ file:///root/.cache/pypoetry/artifacts/8b/9a/dd/79a36c62bc7f37f98d0ea33652570e19272e8a7a2297db13a6785698d1/apache_airflow_providers_ftp-2.0.1-py3-none-any.whl apache-airflow-providers-http @ file:///root/.cache/pypoetry/artifacts/52/28/81/03a89147daf7daceb55f1218189d1c4af01c33c45849b568769ca6765f/apache_airflow_providers_http-2.0.1-py3-none-any.whl apache-airflow-providers-imap @ file:///root/.cache/pypoetry/artifacts/1c/5d/c5/269e8a8098e7017a26a2a376eb3020e1a864775b7ff310ed39e1bd503d/apache_airflow_providers_imap-2.0.1-py3-none-any.whl apache-airflow-providers-postgres @ file:///root/.cache/pypoetry/artifacts/fb/69/ac/e8e25a0f6a4b0daf162c81c9cfdbb164a93bef6bd652c1c00eee6e0815/apache_airflow_providers_postgres-2.3.0-py3-none-any.whl apache-airflow-providers-redis @ file:///root/.cache/pypoetry/artifacts/cf/2b/56/75563b6058fe45b70f93886dd92541e8349918eeea9d70c703816f2639/apache_airflow_providers_redis-2.0.1-py3-none-any.whl apache-airflow-providers-sqlite @ file:///root/.cache/pypoetry/artifacts/61/ba/e9/c0b4b7ef2599dbd902b32afc99f2620d8a616b3072122e90f591de4807/apache_airflow_providers_sqlite-2.0.1-py3-none-any.whl ``` ### Deployment Other Docker-based deployment ### Deployment details AWS ECS + Fargate, RabbitMQ, Celery executor, Postgres 13 backend ### What happened Hey so we just hit an issue with the 2.2.0b2 migration https://github.com/apache/airflow/blob/63b1c270cb09ff897702dafda28c59559f9e6080/airflow/migrations/versions/7b2661a43ba3_taskinstance_keyed_to_dagrun.py#L1 . We have dag_runs without a run_id (as it is still nullable) on the dag_run table. This caused the FK constraint on TI not to apply. Noticeably this wasn't caught by https://github.com/apache/airflow/blob/ca93ff4ef05e68369b4fd42fcda3b40f63682d4a/airflow/utils/db.py#L700 before the migrations apply. Is it worth adding an additional check here and making dag_run.run_id NOT NULL? ### What you expected to happen The data check before applying migrations should have spotted the NULL run_ids on the dag_run table. Going forward perhaps we can make this column NOT NULL? ### How to reproduce NULL some run_id columns on the dag_run table. As to how this occurs I am not sure. ### Anything else _No response_ ### Are you willing to submit PR? - [ ] Yes I am willing to submit a PR! ### Code of Conduct - [X] I agree to follow this project's [Code of Conduct](https://github.com/apache/airflow/blob/main/CODE_OF_CONDUCT.md) -- 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]
