kubeflow / pipelines

Machine Learning Pipelines for Kubeflow
https://www.kubeflow.org/docs/components/pipelines/
Apache License 2.0
3.56k stars 1.61k forks source link

[backend] The backend returns an incorrect pod name to the UI if the AWF POD_NAMES env var is set to "v1" #11019

Open droctothorpe opened 2 months ago

droctothorpe commented 2 months ago

Environment

Steps to reproduce

AWF lets you revert to the v1 pod name syntax using the POD_NAMES env var documented here.

Whether to have pod names contain the template name (v2) or be the node id (v1) - should be set the same for Argo Server.

KFP assumes that POD_NAMES is set to "v2." It should not make this assumption since other applications can depend on the v1 pod name syntax.

To reproduce the error, just set POD_NAMES to "v1" in AWF and try to stream the logs for a task in the KFP GUI.

I'm not sure, but I think we just need to add some extra parsing for a corresponding KFP env var here to make this configurable for end users.

Validating the fix on my end.


Impacted by this bug? Give it a 👍.

github-actions[bot] commented 6 days ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.