I am trying to run socket.io application using Google Container Engine. I have installed a login service that creates a google load balancer that points to the cluster. If I have one block in the cluster everything works well. As soon as I add more I get tons of socket.io errors. It looks like the connections end up in different containers in the cluster and I suspect this is a problem with all socket.io polls and updates.

I am setting up a load balancer to use IP based sticky sessions.

Does this mean that it will have an affinity for a specific NODE in a Kubernetes cluster and not a POD?

How can I configure it to ensure that the session is close to a specific POD in the cluster?

NOTE. I am manually setting sessionAffinity to a cloud load balancer.

This will be my yaml input

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: my-ingress
  annotations:
    kubernetes.io/ingress.global-static-ip-name: my-static-ip
spec:
  backend:
    serviceName: my-service
    servicePort: 80

      

        
        
        
      

    

Service

apiVersion: v1
kind: Service
metadata:
  name: my-service
  labels:
    app: myApp
spec:
  sessionAffinity: ClientIP
  type: NodePort
  ports:
    - port: 80
      targetPort: http-port
  selector:
    app: myApp