Deploy rabbitmq-autocluster on k8s with persistent storage

rabbitmq-autocluster_k8s_persistent.bash

#!/bin/bash

set -eo pipefail

export KUBE_NAMESPACE=test
export REPLICA_COUNT=3

cat <<EOF | kubectl apply -f -
apiVersion: apps/v1beta1
kind: StatefulSet
metadata:
  name: rabbitmq
  namespace: $KUBE_NAMESPACE
spec:
  serviceName: rabbitmq
  replicas: $REPLICA_COUNT
  template:
    metadata:
      labels:
        app: rabbitmq
    spec:
      terminationGracePeriodSeconds: 10
      containers:
      - name: rabbitmq-autocluster
        image: pivotalrabbitmq/rabbitmq-autocluster
        ports:
          - name: http
            protocol: TCP
            containerPort: 15672
          - name: amqp
            protocol: TCP
            containerPort: 5672
        livenessProbe:
          exec:
            command: ["rabbitmqctl", "status"]
          initialDelaySeconds: 30
          timeoutSeconds: 5
        readinessProbe:
          exec:
            command: ["rabbitmqctl", "status"]
          initialDelaySeconds: 10
          timeoutSeconds: 5
        imagePullPolicy: Always
        env:
          - name: MY_POD_IP
            valueFrom:
              fieldRef:
                fieldPath: status.podIP
          - name: NAMESPACE
            valueFrom:
              fieldRef:
                fieldPath: metadata.namespace
          - name: HOSTNAME
            valueFrom:
              fieldRef:
                fieldPath: metadata.name
          - name: RABBITMQ_USE_LONGNAME
            value: "true"
          - name: RABBITMQ_NODENAME
            value: "rabbit@\$(HOSTNAME).rabbitmq.\$(NAMESPACE).svc.cluster.local"
          - name: AUTOCLUSTER_TYPE
            value: "k8s"
          - name: AUTOCLUSTER_DELAY
            value: "10"
          - name: K8S_ADDRESS_TYPE
            value: "hostname"
          - name: K8S_SERVICE_NAME
            value: rabbitmq
          - name: K8S_HOSTNAME_SUFFIX
            value: ".rabbitmq.\$(NAMESPACE).svc.cluster.local"
          - name: AUTOCLUSTER_CLEANUP
            value: "false"
          - name: CLEANUP_WARN_ONLY
            value: "true"
        volumeMounts:
        - name: rabbitmq-data
          mountPath: /var/lib/rabbitmq/mnesia
  volumeClaimTemplates:
  - metadata:
      name: rabbitmq-data
      annotations:
        volume.beta.kubernetes.io/storage-class: gp2
    spec:
      accessModes: [ "ReadWriteOnce" ]
      resources:
        requests:
          storage: 4Gi
EOF

# Headless service for Pod DNS

cat <<EOF | kubectl apply -f -
kind: Service
apiVersion: v1
metadata:
  namespace: $KUBE_NAMESPACE
  name: rabbitmq
  labels:
    app: rabbitmq
spec:
  clusterIP: None
  ports:
   - name: http
     protocol: TCP
     port: 15672
     targetPort: 15672
   - name: amqp
     protocol: TCP
     port: 5672
     targetPort: 5672
  selector:
    app: rabbitmq
EOF

# LoadBalancer service for public access

cat <<EOF | kubectl apply -f -
kind: Service
apiVersion: v1
metadata:
  namespace: $KUBE_NAMESPACE
  name: rabbitmq-lb
  labels:
    app: rabbitmq
    type: LoadBalancer
spec:
  type: LoadBalancer
  ports:
   - name: http
     protocol: TCP
     port: 15672
     targetPort: 15672
   - name: amqp
     protocol: TCP
     port: 5672
     targetPort: 5672
  selector:
    app: rabbitmq
EOF

echo "Waiting for StatefulSet to complete rolled out"
# kubectl rollout status statefulset/rabbitmq -n $KUBE_NAMESPACE # Not supported in k8s v1.6 and prior
for i in $(seq 1 120); do
  if kubectl exec rabbitmq-$(($REPLICA_COUNT - 1)) -n $KUBE_NAMESPACE -- rabbitmqctl status &> /dev/null; then
    break
  fi
  sleep 1s
done
if [[ "$i" == 120 ]]; then
  echo "StatefulSet taking too long to complete, you need to manual install..."
  exit 1
fi

echo "Setting up HA policy"
kubectl exec rabbitmq-0 -n $KUBE_NAMESPACE -- rabbitmqctl set_policy ha-all "" '{"ha-mode":"all","ha-sync-mode":"automatic"}'

• name: AUTOCLUSTER_CLEANUP
  value: "false"

I know this is one way to solve the problem, but this is dangerous. We already try this way, but some time will partition. So will have two clusters. Some client will connet to part one, some will connect part two.
So, it is not my suggestion. If not use this, one is down, but the cluster can work also.