(under development)
This operator is developed with kubebuilder, which can deploy StarRocks CRD resources in kubernetes.

This Kubernetes Operator is able to deploy StarRocks' Front End (FE), Back End (BE) and Compute Node (CN) components into your kubernetes environment. These components run in FQDN (fully qualified domain name) mode by default.

• kubernetes 1.18+
• golang 1.18+

• FE decouples with CN and BE. FE is a must-have component, BE and CN can be optionally deployed.
• Support v2 horizontalpodautoscalers for CN cluster.

Get the official operator image from here.

Follow below instructions if you want to build your own image.

DOCKER_BUILDKIT=1 docker build -t starrocks-kubernetes-operator/operator:<tag> .

E.g.

DOCKER_BUILDKIT=1 docker build -t starrocks-kubernetes-operator/operator:latest .

docker push ghcr.io/OWNER/starrocks-kubernetes-operator/operator:latest

E.g. Publish image to ghcr

docker push ghcr.io/dengliu/starrocks-kubernetes-operator/operator:latest

Apply the custom resource definition (CRD) for the Operator:

kubectl apply -f https://raw.githubusercontent.com/StarRocks/starrocks-kubernetes-operator/main/deploy/starrocks.com_starrocksclusters.yaml

Apply the Operator manifest. By default, the Operator is configured to install in the starrocks namespace. To use the Operator in a custom namespace, download the Operator manifest and edit all instances of namespace: starrocks to specify your custom namespace. Then apply this version of the manifest to the cluster with kubectl apply -f {local-file-path} instead of using the command below.

kubectl apply -f https://raw.githubusercontent.com/StarRocks/starrocks-kubernetes-operator/main/deploy/operator.yaml

You need to prepare a separate yaml file to deploy the StarRocks FE, BE and CN components. The starrocks cluster CRD fields explains in api.md. The examples directory contains some simple example for reference.

You can use any of the template yaml file as a starting point. You can further add more configurations into the template yaml file following this deployment documentation.

Official FE/CN/BE components images can be found from dockerhub:

You can specify the image name in the yaml file. For example, the below configuration uses the starrocks/alpine-fe:2.4.1 image for FE.

starRocksFeSpec:
  image: starrocks/alpine-fe:2.4.1

The official images contains default application configuration file, however, they can be overritten by configuring kubernetes configmap deployment crd.

You can generate the configmap from an StarRocks configuration file. Below is an example of creating a Kubernetes configmap fe-config-map from the fe.conf configuration file. You can do the same with BE and CN.

# create fe-config-map from starrocks/fe/conf/fe.conf file
kubectl create configmap fe-config-map --from-file=starrocks/fe/conf/fe.conf

Once the configmap is created, you can reference the configmap in the yaml file. For example:

# fe use configmap example
starRocksFeSpec:
  configMapInfo:
    configMapName: fe-config-map
    resolveKey: fe.conf
# cn use configmap example
starRocksCnSpec:
  configMapInfo:
    configMapName: cn-config-map
    resolveKey: cn.conf
# be use configmap example
  starRocksBeSpec:
    configMapInfo:
    configMapName: be-config-map
    resolveKey: be.conf

External storage can be used to store FE meta and BE data for persistence. storageVolumes can be specified in corresponding component spec to enable external storage volumes auto provisioning. Note that the specific storageClassName should be available in kubernetes cluster before enabling this storageVolume feature.

If StorageVolume info is not specified in CRD spec, the operator will use emptydir mode to store FE meta and BE data.

FE storage example

starRocksFeSpec:
  storageVolumes:
    - name: fe-meta
      storageClassName: meta-storage
      storageSize: 10Gi
      mountPath: /opt/starrocks/fe/meta # overwrite the default meta path

BE storage example

starRocksBeSpec:
  storageVolumes:
    - name: be-data
      storageClassName: data-storage
      storageSize: 1Ti
      mountPath: /opt/starrocks/be/storage # overwrite the default data path

For demonstration purpose, we use the starrocks-fe-and-be.yaml example template to start a 3 FE and 3 BE StarRocks cluster.

kubectl apply -f starrocks-fe-and-be.yaml

After deploying the StarRocks cluster, you can use kubectl get svc -n <namespace> to find the IP to connect to. For example if the namespace that starrocks is deployed into is starrocks, you can:

kubectl get svc -n starrocks

<your-StarRocksCluster-name>-fe-service's clusterIP is the IP to use to connect to StarRocks FE.

Delete the custom resource:

kubectl delete -f starrocks-fe-and-be.yaml

Remove the Operator:

kubectl delete -f  https://raw.githubusercontent.com/StarRocks/starrocks-kubernetes-operator/main/deploy/operator.yaml

StarRocks have supported helm use.
helm chart. github repo