Is your feature request related to a problem? Please describe.
The one-click quickstart supports a control plane plus GPU cluster layout, and the self-managed stack separates the control-plane lifecycle from the per-cluster NVCA operator lifecycle. We need explicit validation that adding an additional compute cluster after the first install works as expected.
This matters because a self-managed NVCF control plane should support more than one GPU cluster without clobbering registration state or reinstalling the control plane.
Describe the solution you'd like
Validate and document the N+1 cluster flow for one-click deployment.
The validation should cover:
- Install control plane and compute cluster A.
- Add compute cluster B using a separate kube context and cluster name.
- Confirm both clusters remain registered and healthy.
- Confirm functions can deploy to the expected cluster target.
- Confirm uninstalling cluster B does not remove cluster A or the control plane.
Describe alternatives you've considered
Operators can use the manual Helmfile and NVCA registration flow, but one-click is the preferred fresh-install path and should support the common multi-cluster expansion workflow.
Additional context
Suggested acceptance criteria:
- One-click flow succeeds for the first compute cluster.
- One-click or follow-up command succeeds for the second compute cluster.
- Cluster-specific registration values remain isolated.
nvcf-cli self-hosted status shows both clusters correctly.- Docs include the verified N+1 workflow and cleanup commands.
Is your feature request related to a problem? Please describe.
The one-click quickstart supports a control plane plus GPU cluster layout, and the self-managed stack separates the control-plane lifecycle from the per-cluster NVCA operator lifecycle. We need explicit validation that adding an additional compute cluster after the first install works as expected.
This matters because a self-managed NVCF control plane should support more than one GPU cluster without clobbering registration state or reinstalling the control plane.
Describe the solution you'd like
Validate and document the N+1 cluster flow for one-click deployment.
The validation should cover:
Describe alternatives you've considered
Operators can use the manual Helmfile and NVCA registration flow, but one-click is the preferred fresh-install path and should support the common multi-cluster expansion workflow.
Additional context
Suggested acceptance criteria:
nvcf-cli self-hosted statusshows both clusters correctly.