Run Hippo in Minikube

Install virtualbox https://www.virtualbox.org/wiki/Downloads

Install minikube https://github.com/kubernetes/minikube

brew cask install minikube

Start minikube with some additional resources

minikube --memory 8192 --cpus 2 start

Create the persistant volumes and persistant volume claims for pods that require persistence. From kubernetes directory run:

./run-once.sh

Add an entry in your /etc/hosts file. Replace <minikube_ip> with the output of the command:

minikube ip

Add the following line in your /etc/hosts file:

<minikube_ip> backatcha.server hippo.site hippo.cms

Enable ingress addon in minikube

minikube addons enable ingress

To deploy everything at once, from kubernetes directory run:

./deploy-all.sh

Start up the minikube dashboard:

minikube dashboard

When hippo pod(s) have started up, add the following virtual host config so channel manager works: (Considering the site host name is "hippo.site")

/kube:
  jcr:primaryType: hst:virtualhostgroup
  jcr:uuid: 31c0e8e9-c81b-432b-b391-6eaf055cfc13
  hst:cmslocation: http://hippo.cms/cms
  hst:defaultport: 8080
  /site:
    jcr:primaryType: hst:virtualhost
    jcr:uuid: 3c5585e4-ad07-4484-b8cb-189902fbb211
    /hippo:
      jcr:primaryType: hst:virtualhost
      jcr:uuid: 6e40871e-eaf3-47ee-b7bc-3b1b9dd3b6d2
      /hst:root:
        jcr:primaryType: hst:mount
        jcr:uuid: 08a87064-8383-4ba1-a4cf-0ca526f5aac8
        hst:homepage: root
        hst:mountpoint: /hst:hst/hst:sites/myhippoproject
        hst:showcontextpath: false

• You can visit the site at hippo.site/ and the cms at hippo.cms/cms
• You can visit the backatcha spring boot app at backatcha.server/

Local Development without pushing to public docker registry

If you want to do develepment in local projects but not want to push to public image registries, consider the following workflow:

• Make changes in local project, e.g. in hippo's site module.
• Recreate distribution file, e.g. mvn clean verify && mvn -Pdist
• Move distribution file to kubernetes/cms-site folder (remove any existing .tar.gz)
• To be able to work with the docker daemon on your mac/linux host use the docker-env command in your shell

eval $(minikube docker-env)

More info on the above command is at: https://kubernetes.io/docs/setup/minikube#reusing-the-docker-daemon Note that you should keep using the same shell

• Now you can build the docker image and tag it at the same time, from kubernetes/cms-site directory:

docker build username/release-name:version .

You can replace the username, release-name, and version in the command above. Dot(.) refers to the location of the dockerfile which is in kubernetes/cms-site directory.

• At this point the docker image is tagged and pods in the cluster can pull this image. Go to a pod definition file and replace the image. E.g.

#SNIP
 spec:
      containers:
        - name: hippo
          image: username/release-name:version #can pull since we ran the eval command
          ports:
            - containerPort: 8080
#-SNIP

Remote debugger for your hippo project

if you want to debug the deployed hippo project within the cluster:

• Set the environment variable "JAVA_ENABLE_DEBUG" to true in hippo-deployment.yaml
• Uncomment the part that exposes the container port 5005 (in hippo-deployment.yaml)

#SNIP
          ports:
            - containerPort: 8080
#           - containerPort: 5005
#             name: "jvm-debug" #Uncomment these  two lines
#-SNIP

• If you already had a cluster running simply apply the changes (from kubernetes directory):

kubectl apply -f hippo-deployment.yaml

• Activate port forwarding from localhost:5005 to pod's 5005 by running the command:

kubectl port-forward deployments/hippo 5005:5005

• Connect from IntelliJ (or other) remotely to localhost:5005 your breakpoints will be triggered

• If you have more than 1 hippo pod, and you want to trigger breakpoints for only 1 pod port-forward to the specific pod:

kubectl get pods #get the name of the pod you want to set breakpoints for
kubectl port-forward <pod_name_goes_here>  5005:5005

Usage & Resource Monitoring

You can enable heapster https://github.com/kubernetes/heapster on the running cluster by running:

minikube addons enable heapster

Note that this might take a while. Then, open the dashboard by running:

minikube addons open heapster

You can login with "admin/admin" and observe the pods' resource usage.

Remarks

• conf/context.xml has mysql configured. This is typical mysql setup for a Hippo project except that the url for the database server points to:

  mysql.default.svc.cluster.local:3306

  First "mysql" part is the name of the kubernetes service name called mysql. More explanation: https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/

• Pods pull docker images from public docker registry. If you want different docker images you need to build your own and push them to a registry that your local machine has access to

• You can build your own hippo docker image by putting your distribution tar.gz in kubernetes/cms-site directory and running:

docker build -t my-release:latest .

Then you can push it to a public registry and change the images in pod specification yaml files

• Backatcha-server is a simple spring boot app that prints a UUID. Scale up this app and see that the load balancer works.