K8s and Openfire

April 8, 2020

Openfire is a popular choice to implement instant messaging through XMPP.

Thanks to the DevOps cultural shift widely adopted in modern cloud-native software engineering, infrastructure as code is paramount.

Below, I am describing the steps to create a local minikube cluster with instant messaging capabilities via openfire.

Running Minikube #

1. very first time
   • follow the installation macos
   • please make sure that default vm is docker because you want to download images without having to switch docker registries locally:

     minikube config set driver docker
     

2. start minikube

   minikube start
   

3. start minikube dashboard start the dashboard in a separate terminal window because the process should keep running

   minikube dashboard
   

4. create cluster
   • create openfire.yaml with the following content

     apiVersion: v1
     kind: Service
     metadata:
     name: openfire-service
     spec:
     type: NodePort
     selector:
         app: openfire
     ports:
         - protocol: TCP
         port: 9090
         targetPort: 9090
         nodePort: 31090
         name: http
         - protocol: TCP
         port: 5222
         targetPort: 5222
         nodePort: 31222
         name: name2
         - protocol: TCP
         port: 7777
         targetPort: 7777
         nodePort: 31777
         name: name3
     ---
     apiVersion: networking.k8s.io/v1beta1 # for versions before 1.14 use extensions/v1beta1
     kind: Ingress
     metadata:
     name: openfire-ingress
     annotations:
         nginx.ingress.kubernetes.io/rewrite-target: /
     spec:
     backend:
         serviceName: openfire-service
         servicePort: 9090
     rules:
         - host: openfire
         - http:
             paths:
             - path: /
                 backend:
                 serviceName: openfire-service
                 servicePort: 9090
     ---
     apiVersion: apps/v1
     kind: Deployment
     metadata:
     name: openfire-deployment
     labels:
         app: openfire
     spec:
     replicas: 1
     selector:
         matchLabels:
         app: openfire
     template:
         metadata:
         labels:
             app: openfire
         spec:
         containers:
             - name: openfire
             image: sameersbn/openfire:3.10.3-19
             ports:
                 - containerPort: 9090
                 name: http
                 - containerPort: 5222
                 name: name2
                 - containerPort: 7777
                 name: name3
     

   • then apply using kubectl

     kubectl apply -f openfire.yaml
     

5. expose openfire locally
   • List localurls usinh the command below. It will produce a table including the localendpoint that will proxy requests to the openfire-service ordered as 9090, 5222, and 7777

     minikube service openfire-service --url    
     

configure and testing local openfire #

1. configure admin

   • use minikube service openfire-deployment --url to expose the localhost ports.
   • using the first url provided through the command above, open the admin console and configure openfire using the embedded database option (the external database configuration will be covered in fanother post).

2. testing

   • testing can be done using spark; make sure your client points to the local openfire exposed port, second in order, through minikube.

Running in Production Mode #

The real cluster, will require ingress and egress configurations. Those can be achieve with the following instructions.