[GitHub] [airflow] eskarimov edited a comment on issue #18999: Update databricks provider to use TriggerOperator

[GitBox <gi...@apache.org>]

eskarimov edited a comment on issue #18999:
URL: https://github.com/apache/airflow/issues/18999#issuecomment-956445982


   @chinwobble you're right, I see the same behaviour... I've spotted though when I don't pass `timeout` argument to `self.defer()`, a deferred task will fail if its runtime more than `execution_timeout` set for the Operator calling it. 
   I.e. timeout for a deferred task is equal to execution timeout for an Operator in this case. [Here is the code behind](https://github.com/apache/airflow/blob/94a0a0e8ce4d2b54cd6a08301684e299ca3c36cb/airflow/models/taskinstance.py#L1547-L1555)
   
   Also, we might found a a bug, because I see [a code part in `TaskInstance`](https://github.com/apache/airflow/blob/94a0a0e8ce4d2b54cd6a08301684e299ca3c36cb/airflow/models/taskinstance.py#L1495-L1500) trying to check for timeout and re-calculate it, when task executed after deferred state. The issue is that `self.start_date` is changing every time when a task is continued after the deferred state.


-- 
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: commits-unsubscribe@airflow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org