Stop, Start, and Restart a ClickHouse Cluster

Prerequisites

• A functional Kubernetes cluster (v1.21+ recommended)
• kubectl v1.21+ installed and configured with cluster access
• KubeBlocks installed (installation guide)
• ClickHouse Add-on enabled
• A demo namespace: kubectl create ns demo

Stop the Cluster

Stopping a cluster suspends all pods while retaining PVCs and data. Useful for saving compute costs during idle periods.

kubectl apply -f - <<EOF
apiVersion: operations.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
  name: clickhouse-stop
  namespace: demo
spec:
  clusterName: clickhouse-cluster
  type: Stop
EOF

Monitor progress:

kubectl get cluster clickhouse-cluster -n demo -w

NAME                  CLUSTER-DEFINITION   TERMINATION-POLICY   STATUS     AGE
clickhouse-cluster    clickhouse           Delete               Stopping   10s
clickhouse-cluster    clickhouse           Delete               Stopped    45s

Start the Cluster

kubectl apply -f - <<EOF
apiVersion: operations.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
  name: clickhouse-start
  namespace: demo
spec:
  clusterName: clickhouse-cluster
  type: Start
EOF

NAME                  CLUSTER-DEFINITION   TERMINATION-POLICY   STATUS    AGE
clickhouse-cluster    clickhouse           Delete               Running   2m

Restart a Component

Restart the ClickHouse shard component:

kubectl apply -f - <<EOF
apiVersion: operations.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
  name: clickhouse-restart
  namespace: demo
spec:
  clusterName: clickhouse-cluster
  type: Restart
  restart:
    - componentName: clickhouse
EOF

For clusters with Keeper, restart the Keeper component:

kubectl apply -f - <<EOF
apiVersion: operations.kubeblocks.io/v1alpha1
kind: OpsRequest
metadata:
  name: clickhouse-restart-keeper
  namespace: demo
spec:
  clusterName: clickhouse-cluster
  type: Restart
  restart:
    - componentName: ch-keeper
EOF

Monitor the OpsRequest:

kubectl get opsrequest -n demo -w

NAME                  TYPE      CLUSTER              STATUS    PROGRESS   AGE
clickhouse-restart    Restart   clickhouse-cluster   Succeed   1/1        30s

Cleanup

kubectl delete opsrequest clickhouse-stop clickhouse-start clickhouse-restart clickhouse-restart-keeper -n demo --ignore-not-found