apecloud/kubeblocks-create-cluster

Create a Database Cluster on KubeBlocks

This skill is the other-addons fallback.

Use it only when:

• the target engine does not have a dedicated addon-* entry skill
• KubeBlocks is already installed
• preflight has already produced storage and routing recommendations

Do not use this skill for:

• MySQL -> kubeblocks-addon-mysql
• PostgreSQL -> kubeblocks-addon-postgresql
• Redis -> kubeblocks-addon-redis
• MongoDB -> kubeblocks-addon-mongodb
• Kafka -> kubeblocks-addon-kafka

If rollout readiness is still unknown, stop and run kubeblocks-preflight first.

Workflow

- [ ] Step 1: Confirm this is truly an other-addons case
- [ ] Step 2: Gather clusterDef / topology / component names from the environment
- [ ] Step 3: Fill the generic Cluster template
- [ ] Step 4: Validate and apply
- [ ] Step 5: Verify the cluster becomes Running

Step 1: Confirm This Is Truly An Other-Addons Case

Before using the generic path, answer both questions:

1. Does the engine already have a dedicated addon-* skill in this repository?
2. Has preflight already selected storage, sizing, and namespace defaults?

If the answer to either question is "no", stop and route correctly first.

Step 2: Gather Engine-Specific Values

Use the live cluster to discover the needed values:

kubectl get clusterdefinitions
kubectl get componentversions

Record:

• clusterDef
• topology
• component name
• serviceVersion
• storageClassName

Step 3: Fill The Generic Template

This is a placeholder template for engines that do not have a dedicated entry skill.

apiVersion: apps.kubeblocks.io/v1
kind: Cluster
metadata:
  name: <cluster-name>
  namespace: <namespace>
spec:
  clusterDef: <clusterDef-from-environment>
  topology: <topology-from-environment>
  terminationPolicy: Delete
  componentSpecs:
    - name: <component-name>
      serviceVersion: "<serviceVersion>"
      replicas: <replica-count>
      resources:
        requests:
          cpu: "<cpu-request>"
          memory: "<memory-request>"
        limits:
          cpu: "<cpu-limit>"
          memory: "<memory-limit>"
      volumeClaimTemplates:
        - name: data
          spec:
            accessModes: [ReadWriteOnce]
            storageClassName: <storageClassName-from-preflight>
            resources:
              requests:
                storage: <storage-size>

This template is intentionally generic. It does not try to encode the topology advice for high-frequency engines.

Step 4: Validate And Apply

kubectl apply -f cluster.yaml --dry-run=server
kubectl apply -f cluster.yaml

Step 5: Verify

kubectl get cluster -n <namespace> <cluster-name> -w
kubectl get pods -n <namespace> -l app.kubernetes.io/instance=<cluster-name>

Exit Criteria

This path is complete only when:

• the engine truly belongs to other-addons
• storage choice comes from preflight, not guesswork
• the cluster reaches Running

For connection examples and deep engine-specific operations, rely on the relevant addon skill or official addon docs instead of expanding this fallback into another high-frequency entry path.