Deploy application on VM

In the previous chapter we deployed the customer and preference service on to cluster1 Kubernetes cluster. To complete our demo services interaction let us deploy recommendation on the VM.

As there are lot of moving parts¹ in doing this setup, I packed them into set of Ansible tasks. When you run the tasks of the playbook, it will do all that is required setup Istio Sidecar and recommendation namely,

• Recommendation v1 service running in vm as daemon
• Istio Sidecar installed and configured on the VM
• IP Routes from VM to Kubernetes Service and Pods. Check Istio VM Prerequisites for more details.
• Create a gateway so that the VM Istio Sidecar (Envoy) can communicate securely (mTLS) with istiod on cluster1.

At end of this chapter we should be able to make the service interaction flow work as shown below.

Demo Service Interactions

Run the following command to prepare the VM with required packages,

make deploy-base

Then running the following command will make the VM ready with recommendation app with Istio Sidecar,

make deploy-workload

Resources

Pods and Services

kubectl --context="$CLUSTER1" get svc,pods

NAME                             READY   STATUS    RESTARTS   AGE
pod/customer-6bbb45d7db-5db6l    2/2     Running   0          56m
pod/preference-b95d64c99-9wjfp   2/2     Running   0          46m

NAME                     TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)    AGE
service/kubernetes       ClusterIP   172.18.0.1    <none>        443/TCP    40h
service/customer         ClusterIP   172.18.7.18   <none>        8080/TCP   75m
service/preference       ClusterIP   172.18.6.50   <none>        8080/TCP   46m
service/recommendation   ClusterIP   172.18.6.63   <none>        8080/TCP   2m19s

Workload Entry

A recommendation Workload Entry,

kubectl --context="$CLUSTER1" workloadentries.networking.istio.io -owide

Name:         recommendation-vm1
Namespace:    default
Labels:       <none>
Annotations:  <none>
API Version:  networking.istio.io/v1beta1
Kind:         WorkloadEntry
Spec:
  Address:  192.168.64.92
  Labels:
    App:            recommendation
    instance_id:    vm1
    Version:        v1
  Network:          network1
  Service Account:  vm-service-account
Events:             <none>

As you noticed in the above output the output the address of the WorkloadEntry points to the vm1 IP address. The recommendation service will use the selector labels from the WorkloadEntry i.e App to route the traffic to the vm1 using the IP Address from WorkloadEntry.

Note

You can get vm1 details using the command multipass info vm1

Interacting with Services

Retrieve the Istio Ingress Gateway url to access the application,

export INGRESS_GATEWAY_IP=$(kubectl --context ${CLUSTER1} -n istio-gateways get svc ingressgateway -o jsonpath='{.status.loadBalancer.ingress[0].*}')
export SVC_URL="${INGRESS_GATEWAY_IP}/customer"

Call Service

Call the service using the script,

$DEMO_HOME/bin/call_service.sh

Poll Service

Poll the service using the script,

$DEMO_HOME/bin/poll_service.sh

Using Default Browser

Open the URL in the browser open http://$SVC_URL.

The customer service should now succeed completing the call to recommendation service on the vm1,

customer => preference => recommendation v1 from 'vm1': 1

Voilà! We have now successfully connected all the services right from Kubernetes to VM.

The communication from vm to k8s should also work, lets do a quick test calling same customer service,

multipass exec vm1 -- curl customer.default.svc.cluster.local:8080

And you should get pretty much the same response,

customer => preference => recommendation v1 from 'vm1': 2

In next chapter we will see how to apply various Service Mesh features like Rate Limit, Access Policies etc., using Gloo Mesh.

---

1. https://istio.io/v1.11/docs/setup/install/virtual-machine ↩