GitHub user potiuk added a comment to the discussion: Inconsistent Database 
Resolution Order `sql_alchemy_conn` vs `sql_alchemy_conn_secret`

I seriously doubt this is a problem Airflow uses the same resolution mechanism 
in consistent order.

Most likely what was different is that you had DIFFERENT configuration in two 
places where things were running. For example in worker you did not have SECRET 
configured. This is impossible expectation to use a configuration that you have 
no access to.

Generally speaking the expectation is that it's on yoy (deployment manager) to 
make sure your configuratoin is consistently set on all your nodes and in all 
processes of Airflow. That's very basic assumption of airflow environment 
documented in 
https://airflow.apache.org/docs/apache-airflow/stable/configurations-ref.html#configuration-reference

> Use the same configuration across all the Airflow components. While each 
> component does not require all, some configurations need to be same otherwise 
> they would not work as expected. A good example for that is 
> [secret_key](https://airflow.apache.org/docs/apache-airflow/stable/configurations-ref.html#config-webserver-secret-key)
>  which should be same on the Webserver and Worker to allow Webserver to fetch 
> logs from Worker.


Note that this is changing in Airflow 3, because the "workers" will no longer 
have database access and they wil not require most of configuration so the 
configuration for sql_alchemy_conn will only be needed in scheduler, webserver 
and triggerer (but it will still need to be consistent across all three 
components)

Converting it into discussion in case more concrete evidences of the error and 
inconsistency is on airflow side (but it's practically impossible).

GitHub link: 
https://github.com/apache/airflow/discussions/44799#discussioncomment-11512421

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to