-
|
Hi All! I'm designing an Airflow platform in an enterprise environment. It runs on OpenShift, within containers. Everything was worked fine, but we had to start to migrate to an EDB database, outside Kubernetes. This is OK, EDB is just a "fancy" Postgres. We made our compatibility test, and everything went well. I already integrated EDB in our Sandbox environment. It works just fine since 2 weeks. Without any issue. Airflow is in the latest release btw. Yesterday I started the EDB integaration of our next environment. This EDB is a completely different instance (different DB, VM, everything). And I cannot tell why, but all of the Airflow components (scheduler, triggerer, webserver) are constantly crashing with similar stacktrace like this: File "/home/airflow/.local/lib/python3.8/site-packages/airflow/jobs/triggerer_job.py", line 106, in _execute It is clear, that this issue has to do something with the Database and the heartbeat, but I can't figure it out exactly what. I asked the DBAs to differentiate the settings of the two databases, but they found nothing. Is there anyone who maybe familiar with this issue and can help me? Thanks in advance! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
Okay, now we figured out, The datestyle of the database was the issue. It was setted something that represented a date like this: 21-OCT-22. This was the issue with the month. The solution was to adjust the database's datesytle to ISO. |
Beta Was this translation helpful? Give feedback.
Okay, now we figured out, The datestyle of the database was the issue. It was setted something that represented a date like this: 21-OCT-22. This was the issue with the month. The solution was to adjust the database's datesytle to ISO.