jigs031 opened a new issue, #32485:
URL: https://github.com/apache/airflow/issues/32485

   ### Apache Airflow version
   
   Other Airflow 2 version (please specify below)
   
   ### What happened
   
   We have spined up the Airflow from Helm chart 1.6.0. We have defined 
AirflowExcecutor as "KubernetesExecutor" in our config map.
   
   System works fine when number of parallel runs are less than, but when 
parallel runs increase - few of the PODS randomly runs on "LocalExecutor" and 
these task are not visible on Airflow DAG Run UI screen. After the long 
debugging we identified that it was dues to the environment variable 
"AIRFLOW__CORE__EXECUTOR " hard-coded to "LocalExecutor" in 
chart/files/pod-template-file.kubernetes-helm-yaml.
   
   We tried to hard-coded as "KubernetesExecutor", then all the Jobs are 
working as expected.
   
   ### What you think should happen instead
   
   Like other configuration, Variable AIRFLOW__CORE__EXECUTOR in the pod 
template file should not be hardcoded (there is a possibility to add it via 
.Values.<variable_name>).
   
   ### How to reproduce
   
   Install the helm chart and define - "AIRFLOW__CORE__EXECUTOR == 
"KubernetesExecutor" in ConfigMap.
   
   Run single job multiple times (atleast 20+ runs) within 10 seconds. You may 
observe that - not all job instances are visible on the screen. If we go on 
Task level details - you may find task are available but the Job are not 
register on Airflow UI/Database.
   
   ### Operating System
   
   Linux
   
   ### Versions of Apache Airflow Providers
   
   2.3.1
   
   ### Deployment
   
   Other 3rd-party Helm chart
   
   ### Deployment details
   
   Official Helm Chart version: 1.6.0
   
   ### 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]

Reply via email to