Manually Creating a TcaNodePool (tknp)

Products

• VMware Telco Cloud Automation

Issue/Introduction

Procedure for manually creating a TcaNodePool (tknp)

Environment

3.X

Resolution

Step 1: Generate a Template (Optional but Recommended)

Since the TcaNodePool schema is complex, it is safest to export an existing node pool configuration to use as a template.

1.  Find an existing node pool in your target namespace:
   
    kubectl get tknp -n <your-namespace>
   
   
2. Export it to a YAML file:
   
   kubectl get tknp <existing-nodepool-name> -n <your-namespace> -o yaml > new-nodepool.yaml
3. Clean up the YAML file:
   
   Remove metadata.uid, metadata.resourceVersion, metadata.creationTimestamp, metadata.generation, and the status section.
   
   

Step 2: Define the TcaNodePool Manifest

 

Create or edit your YAML file (e.g., new-tknp.yaml). Below is a structural example of what the TcaNodePool resource looks like. Adjust the replicas, memory, cpu, and network settings to match your requirements. 

 

YAML

----

apiVersion: telco.vmware.com/v1

kind: TcaNodePool

metadata:

   name: <new-nodepool-name>    # e.g., worker-pool-gpu-01

 

namespace: <target-namespace> # The namespace where the Cluster exists

   labels:

      cluster.x-k8s.io/cluster-name: <cluster-name> # Must match the Cluster name

spec:

    clusterName: <cluster-name> # Link to the parent Cluster

    replicas: 3 # Number of worker nodes

 

# Infrastructure definition (simplified example)

nodeConfig:

   memory: 16384 # RAM in MB

   cpu: 4 # vCPUs

   storage: 50 # Disk in GB

 

# Network settings (crucial for TCA)

networks:

   - name: <network-name> # e.g., "sriov-network-1"

   label: <network-label>

 

# Kubernetes settings

k8sConfig:

   version: <k8s-version> # e.g., v1.20.5+vmware.1

 

Note: The exact fields under spec (like vsphere, networks, or customizations) depend heavily on your underlying infrastructure (vSphere, VCD, OpenStack) and TCA version. Always refer to an existing tknp in your environment for the correct schema.

 

 

Step 3: Apply the Manifest

 

Once your YAML file is ready, apply it to the cluster:

 

kubectl apply -f new-tknp.yaml

 

Step 4: Verify Creation

 

Monitor the status of the new node pool.

 

1. Check if the resource was accepted: 

2. kubectl get tknp -n <namespace>

3. Watch the associated Machine deployments. TCA (via the Cluster API) will begin creating the actual virtual machines (Machines) corresponding to your new pool: 

4. kubectl get machines -n <namespace> -l

cluster.x-k8s.io/deployment-name=<new-nodepool-name>

 

 

 

Additional Information

Troubleshooting

 

- Status "NotReady": If the tknp stays in a pending state, check the logs of the TCA controllers:

   kubectl logs -n tca-system -l app=tca-controller-manager

 

- Validation Errors: If kubectl apply fails, ensure your cluster.x-k8s.io/cluster-name label exactly matches the parent cluster name.