WARNING - Model file does not exist #32043
Replies: 3 comments 9 replies
-
What do you mean by works as expected? does it load the configurations from the |
Beta Was this translation helpful? Give feedback.
-
I mean that the pods do get spun up anyway, and tasks are executed as expected. It's just that this warning appears first and doesn't make sense. I think despite the warning it works because of: |
Beta Was this translation helpful? Give feedback.
-
The file for sure does not exist:
The question is where and for whom - look at where the log file is generated and check there. My bet is that the file does not exist on scheduler, where it should be present and it should be accessible to the unix user that is used by Airflow. So you need to check there. You might have also similar problems if for example the file or the folder where it is located do not have the right permissions for the unix uses that tries to read the file - this is how linux work, if you do not have permissions to read a folder where the file is located, the OS will report the file as not existing. So whatever you do as "different" linux user, might not matter because your airflow process might be run using a differeint linux user. You need to check using the same environment your software is running with. Converted to discussion in case you need more help with your investigation, but I suggest you look deeper. |
Beta Was this translation helpful? Give feedback.
-
Apache Airflow version
Other Airflow 2 version (please specify below)
What happened
Airflow v2.6.0 w/KubernetesExecutor
At this point, it is just a cosmetic issue since everything works as expected. I'm seeing this warning on all of my tasks:
[2023-06-20, 11:40:18 UTC] {pod_generator.py:529} WARNING - Model file /opt/airflow/pod_templates/pod_template_file.yaml does not exist
But the file does exist, as shown in the warning message as it managed to get the path from
kube_config
and I can use yaml.safe_load on it if I were to shell into the pod, or importdeserialize_model_file
from pod_generator.py. What could be going on here?What you think should happen instead
The warning should not be printed as the file does exist.
How to reproduce
Just execute any task.
Operating System
Debian GNU/Linux
Versions of Apache Airflow Providers
No response
Deployment
Official Apache Airflow Helm Chart
Deployment details
Deployed on AKS with official helm chart.
Anything else
No response
Are you willing to submit PR?
Code of Conduct
Beta Was this translation helpful? Give feedback.
All reactions