potiuk commented on issue #16988:
URL: https://github.com/apache/airflow/issues/16988#issuecomment-886168521


   This is normal behaviour when there is a problem with database that 
scheduler (or any other component might crash in case of the database failure. 
There are some other problems in the log you provided (`Kubernetes resource 
version is too old, must reset to 0`) which might indicate that there is a 
problem with your K8S deployment. 
   
   I assume what happened here (please double check and let us know if it did 
not).
   
   If Airflow crashed (which is expect in case of DB failure after several 
internal retries by the SQLAlchemy - see #15496) and your deployment did not 
restart it, then you need to fix the deployment to restart airflow scheduler in 
case of failure. 
   
   The problem could be because of the way the Bitnami chart is implemented. We 
are not controlling the bitnami Helm chart (which I believe you are using), so 
I am not sure if it properly restarts scheduler in case of failure - you will 
need to double check yourself.
   
   Alternatively I recommend you to switch to the official Helm chart relased 
by the Apache Airflow project (available at 
https://airflow.apache.org/docs/helm-chart/stable/index.html) - it's done in 
the way that it will restart scheduler if it crashes and genarally supports 
wide number of options (and we will be able to support you here if you have 
problems with it. 
   
   Closing it, unless there are more information indicating that this is an 
airflow, rather than deployment problem.


-- 
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]


Reply via email to