Description
/kind bug
What steps did you take and what happened:
[A clear and concise description of what the bug is.]
When deploying notebook-controller via v1.6.0-rc.0
the manager
container of notebook-controller
failed to start to due to missing permissions to get,create,patch
coordination.k8s.io
leases
for the leader-election resource. I got past the issue by creating a patch in my manifests repo to add the leader-election permissions needed.
What did you expect to happen:
The leader-election role found here, should have the below config added to prevent notebook controller from failing to start if leader election gets enabled.
apiGroups:
- coordination.k8s.io
resources:
- leases
verbs:
- get
- list
- create
- update
- delete
- patch
- watch
Anything else you would like to add:
I think this issue came about updating the notebook-controller go version to 1.17
as this was not an issue before in kubeflow v1.5.0
.
Environment:
- Kubeflow version: (version number can be found at the bottom left corner of the Kubeflow dashboard):
v.1.6.0-rc.0
- Kubernetes platform: (e.g.
minikube
) eks - Kubernetes version: (use
kubectl version
): 1.21
Metadata
Metadata
Assignees
Type
Projects
Status
Status