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
One would expect that by using kfp.kubernetes.CreatePVC without specifying storage_class_name results in using this default storage class. Instead, a hardcoded standard storage class is used (no matter if such class even exist on the cluster)
Environment
How did you deploy Kubeflow Pipelines (KFP)? - as part of Red Hat OpenShift AI
KFP version: 2.2.0
KFP SDK version: 2.10.1
Steps to reproduce
Use kfp.kubernetes.CreatePVC task without storage_class_name. A PVC with .spec.storageClassName: "standard" is created instead of one using the Storage class actually set as the default one in Kubernetes.
Expected result
Not specifying storage_class_name results in submitting the PVC without .spec.storageClassName property set and therefore result in the Kubernetes default storage class being applied.
Kubernetes has a concept of a default Storage Class.
One would expect that by using
kfp.kubernetes.CreatePVC
without specifyingstorage_class_name
results in using this default storage class. Instead, a hardcodedstandard
storage class is used (no matter if such class even exist on the cluster)Environment
Steps to reproduce
Use
kfp.kubernetes.CreatePVC
task withoutstorage_class_name
. A PVC with.spec.storageClassName: "standard"
is created instead of one using the Storage class actually set as the default one in Kubernetes.Expected result
Not specifying
storage_class_name
results in submitting the PVC without.spec.storageClassName
property set and therefore result in the Kubernetes default storage class being applied.Materials and Reference
pipelines/backend/src/v2/driver/driver.go
Lines 1695 to 1703 in 2e05b3d
Impacted by this bug? Give it a 👍.
The text was updated successfully, but these errors were encountered: