You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thank you for creating this issue. You are right.
@kubeflow/wg-training-leads Do you know the reason why only PyTorrchJob and PaddlePaddleJob use OnFailure as a default RestartPolicy?
In manifests/base/crds/kubeflow.org_pytorchjobs.yaml, it is mentioned that:
description: Restart policy for all replicas within the job. One of Always, OnFailure, Never and ExitCode. Default to Never.
However, in the code of pkg/apis/kubeflow.org/v1/pytorch_types.go, the DefaultRestartPolicy is defined as OnFailure.
PyTorchJobDefaultRestartPolicy is default RestartPolicy for PyTorchReplicaSpec. PyTorchJobDefaultRestartPolicy = RestartPolicyOnFailure
This has caused some confusion. It is recommended that the true default value of RestartPolicy be clarified.
The text was updated successfully, but these errors were encountered: