Hawk Container Edition Components YAML Files

Hawk Cluster Manager YAML File for AWS

The Hawk Cluster Manager (hkce_clustermanager) is a seed node to which all other containers connect. Thus, this container should be reachable from all the containers. Also the network information should not be lost even when it is restarted or rescheduled.

Use the Kubernetes StatefulSets for deploying Hawk Cluster Manager. Like a Deployment, a StatefulSet manages Pods that are based on an identical container specifications. Unlike a Deployment, a StatefulSet maintains a sticky identity for each of their Pods. For details about Kubernetes StatefulSets, see the Kubernetes documentation. Also, create a headless service so that all other containers can reach Hawk Cluster Manager container. In a headless service you specify the clusterIP as None . For details about headless service, see to the Kubernetes documentation. For the Pods created by StatefulSets with headless service, DNS entries are created in the form <pod-name>.<headless-service-name>.

For example, for a StatefulSet with N replicas, each Pod in the StatefulSet is assigned an integer ordinal, from 0 up through N-1, that is unique over the Set. So if your StatefulSets has a name "my-stateful-set" with 3 replicas, it creates three pods with names my-stateful-set-0, my-stateful-set-1 and my-stateful-set-2. Also, if you create a headless service with the name "my-service", then you can connect to these StatefulSets pods using the host name my-stateful-set-0.my-service, my-stateful-set-1.my-service, and so on.

For the Hawk Cluster Manager, create only one hkce_clustermanager pod using StatefulSets with the name hkce-clustermanager-set and headless service with the name hkce-clustermanager-service . Thus, other components can connect to this pod using the hostname hkce-clustermanager-set-0.hkce-clustermanager-service. Set the clusterIP as None for the headless service.

Set up the container (the containers tag) with the Docker image (image) of the Hawk Cluster Manager.

Sample Content for Hawk Cluster Manager YAML File

apiVersion: v1
kind: Service
metadata:
  name: hkce-service
spec:
  ports:
  - port: 2561
    protocol: TCP
    targetPort: 2561
  selector:
    app: hkce-clustermanager
  clusterIP: None
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  labels:
    app: hkce-clustermanager
  name: hkce-clustermanager-set
spec:
  serviceName: hkce-service
  selector:
    matchLabels:
      app: hkce-clustermanager
  template:
    metadata:
      labels:
        app: hkce-clustermanager
    spec:
      containers:
      - name: hkce-clustermanager
        image: <aws_docker_registry>/hkce_clustermanager:2.0
        imagePullPolicy: Always
        env:
        - name: tcp_self_url
          value: hkce-clustermanager-set-0.hkce-service:2561
        - name: tcp_daemon_url
          value: hkce-clustermanager-set-0.hkce-service:2561
        ports:
        - containerPort: 2561
          protocol: TCP

Hawk Agent YAML File for AWS Deployment

Hawk agent must be deployed on all nodes. Thus, for Hawk agent the resource type should be DaemonSet. A DaemonSet ensures that all (or some) nodes run a copy of a Pod.

Set up the container (the containers tag) with the Docker image (image) of the Hawk agent. The tcp_self_url should be set to the hostname of the pod in which Hawk agent container is running. You can use the Kubernetes downward API to get the hostname of the pod. For this set fieldPath to status.podIP . This ensures that each hkce_agent has a unique tcp_self_url. Also the ami_tcp_session for each agent should be set to the node IP on which this hkce_agent container is deployed. You can use the Kubernetes downward API to get the hostname of the node. For this set fieldPath to status.hostIP. The AMI applications running on that node can use same status.hostIP in their daemon_url. Specify the protocol and port to connect to this service.

Sample Content for Hawk Agent YAML File

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: hkce-agent-set
  labels:
    app: hkce-agent
spec:
  selector:
    matchLabels:
      name: hkce-agent-set
  template:
    metadata:
      labels:
        name: hkce-agent-set
    spec:
      containers:
      - name: hkce-agent
        image: <aws_docker_registry>/hkce_agent:2.0
        imagePullPolicy: Always
        env:
        - name: HOST_NAME
          valueFrom:
            fieldRef:
              apiVersion: v1
              fieldPath: status.podIP
        - name: HOST_IP
          valueFrom:
            fieldRef:
              fieldPath: status.hostIP
        - name: tcp_self_url
          value: ${HOST_NAME}:2551
        - name: tcp_daemon_url
          value: hkce-clustermanager-set-0.hkce-service:2561
        - name: ami_tcp_session
          value: ${HOST_IP}:2571
        - name: config_path
          value: /tibco.home/hkce/2.0/config/
        - name: DOCKER_HOST
          value: unix:///var/run/docker.sock
        volumeMounts:
        - mountPath: /var/run/docker.sock
          name: docker-sock-volume
        ports:
        - containerPort: 2551
          name: agentport
          protocol: TCP
        - containerPort: 2571
          hostPort: 2571
          protocol: TCP
      volumes:
      - name: docker-sock-volume
        hostPath:
         path: /var/run/docker.sock

Hawk Console YAML File for AWS Deployment

Setup the a pod with load-balancer service for deploying Hawk Console. The load balancer provides an externally-accessible IP address that sends traffic to the correct port on your cluster nodes.

The load-balancer service would enable us to access the console 8083 port from outside. Set up the container ( the containers tag) with the Docker image (the image tag) of the Hawk Console. Also like Hawk agent, the Hawk Console uses the Kubernetes downward API to access hostname of the pod. For this set fieldPath of HOST_NAME to status.podIP and use this in the tcp_self_url.

Sample Content for Hawk Console YAML File

apiVersion: v1
kind: Service
metadata:
  name: hkce-console-service
spec:
  type: LoadBalancer 
  ports:
  - port: 8083
    targetPort: 8083
  selector:
   app: hkce-console
---

apiVersion: v1
kind: Pod
metadata:
  name: hkce-console
  labels:
    name: hkce-console 
    app: hkce-console 
spec:
  containers:
  - name: hkce-console 
    image: <aws_docker_registry>/hkce_console:2.0
    imagePullPolicy: Always
    env:
    - name: HOST_NAME
      valueFrom:
       fieldRef:
         apiVersion: v1
         fieldPath: status.podIP
    - name: tcp_self_url
      value: ${HOST_NAME}:2551
    - name: tcp_daemon_url
      value: hkce-clustermanager-set-0.hkce-service:2561
    - name: hawk_domain
      value: default
    - name: hawk_console_repository_path
      value: /tibco.home/hkce/2.0/repo
    ports:
    - containerPort: 2551
      name: consoleport
      protocol: TCP
    - containerPort: 8083