Kubernetes and Crate

Author
Claus Matzinger
Filed under
Date
August 16, 2016

During our hackathon last weekend we came up with a template that would easily let users spin up a cluster with Kubernetes. Based on one of our users' work, we improved some details and in order to get a three-node testing cluster up and running you can use this:

apiVersion: v1
kind: Service
metadata:
  name: crate-discovery
  labels:
    app: crate
spec:
  ports:
  - port: 4200
    name: crate-web
  - port: 4300
    name: cluster
  type: LoadBalancer
  selector:
    app: crate
---
apiVersion: "apps/v1alpha1"
kind: PetSet
metadata:
  name: crate
spec:
  serviceName: "crate-db"
  replicas: 3
  template:
    metadata:
      labels:
        app: crate
      annotations:
        pod.alpha.kubernetes.io/initialized: "true"
    spec:
      containers:
      - name: crate
        image: crate:latest
        command:
          - /docker-entrypoint.sh
          - -Des.cluster.name=${CLUSTER_NAME}
          - -Des.multicast.enabled=false
          - -Des.discovery.type=srv
          - -Des.discovery.zen.minimum_master_nodes=2
          - -Des.gateway.recover_after_nodes=2
          - -Des.gateway.expected_nodes=${EXPECTED_NODES}
          - -Des.discovery.srv.query=_cluster._tcp.crate-discovery.default.svc.cluster.local
        volumeMounts:
            - mountPath: /data
              name: data
        resources:
          limits:
            memory: 2Gi
        ports:
        - containerPort: 4200
          name: db
        - containerPort: 4300
          name: cluster
        env:
        # Half the available memory.
        - name: CRATE_HEAP_SIZE
          value: "1g"
        - name: EXPECTED_NODES
          value: "3"
        - name: CLUSTER_NAME
          value: "my-crate-cluster"
      volumes:
        - name: data
          hostPath:
            emptyDir:
              medium: "Memory"

This has certainly some trade-offs, and you can read all about them in our official documentation. Specifically, pay attention to storage, which should be suitable for high-availability data retention. Furthermore, this is a work in progress and will be updated as soon as there's new stuff that makes Crate and Kubernetes work even better :)

Tell us what you think!

We would love some feedback on the template if it helped your cluster management and what we could do better in order to help your use case(s)!

Back to topAll posts