-
Bug
-
Resolution: Unresolved
-
Undefined
-
None
-
4.17.z
-
Quality / Stability / Reliability
-
False
-
-
None
-
Critical
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
What we're missing is endpoint slice management when using a secondary network, without the default pod network and with a custom DNS basedomain. So the base pattern, if we were to do this manually; Create EndpointSlice on management cluster with all of the NodePool Workers + http/https NodePort Create a LoadBalancer Service that grabs an IP from MetalLB and maps it to the custom EndpointSlice via `kubernetes.io/service-name` We need a controller to manage the EPS for NodePool scaleup / scaledown events
Version-Release number of selected component (if applicable):
4.17.x
How reproducible:
Hosted Cluster with a secondary network and the default pod network disabled using a customer base domain.
Actual results:
Need to done manually and can not tolerate node changes.
Expected results:
Additional info: