K
Q

Kubernetes Multiple Service Conflict

12/5/2018

I'm newbie in Kubernetes. I created a Kubernetes Cluster on Amazon EKS. I'm trying to setup multiple kubernetes services to run multiple ASP.NET applications in one cluster. But facing a weird problem.

Everything runs fine when there is only 1 service. But whenever i create 2nd service for 2nd application it creates a conflict. The conflict is sometime service 1 url load service 2 application and sometime it loads service 1 application and same happens with service 2 url on simple page reload.

I've tried both Amazon Classic ELB (With LoadBalancer Service Type) and Nginx Ingress controller (With ClusterIp Service Type). This error is common in both approaches.

Both services and deployments are running on port 80, I even tried different ports for both services and deployments to avoid port conflict but same problem.

I've checked the deployment & service status, and pod log everything looks fine. No error or warning at all

Please guide how i can fix this error. Here is the yaml file of both services for nginx ingress

# Service 1 for deployment 1 (container port: 1120)
apiVersion: v1
kind: Service
metadata:
  creationTimestamp: 2018-12-05T14:54:21Z
  labels:
    run: load-balancer-example
  name: app1-svc
  namespace: default
  resourceVersion: "463919"
  selfLink: /api/v1/namespaces/default/services/app1-svc
  uid: a*****-****-****-****-**********c
spec:
  clusterIP: 10.100.102.224
  ports:
  - port: 1120
    protocol: TCP
    targetPort: 1120
  selector:
    run: load-balancer-example
  sessionAffinity: None
  type: ClusterIP
status:
  loadBalancer: {}

2nd Service

# Service 2 for deployment 2 (container port: 80)
apiVersion: v1
kind: Service
metadata:
  creationTimestamp: 2018-12-05T10:13:33Z
  labels:
    run: load-balancer-example
  name: app2-svc
  namespace: default
  resourceVersion: "437188"
  selfLink: /api/v1/namespaces/default/services/app2-svc
  uid: 6******-****-****-****-************0
spec:
  clusterIP: 10.100.65.46
  ports:
  - port: 80
    protocol: TCP
    targetPort: 80
  selector:
    run: load-balancer-example
  sessionAffinity: None
  type: ClusterIP
status:
  loadBalancer: {}

Thanks

-- Osama Sheikh
amazon-eks
amazon-web-services
asp.net-core
kubernetes
kubernetes-ingress

Similar Questions

Kubernetes service on windows node not reachable
do replicas decrease traffic on a single node kubernetes cluster?
How to delete a daemonset and service in Kubernates
Kubernetes : Replication Controller still there after deletion

1 Answer

12/5/2018

The problem is with the selector in the services. They both have the same selector and that's why you are facing that problem. So they both will point to same set of pods.

The set of Pods targeted by a Service is (usually) determined by a Label Selector

Since deployemnt 1 and deployment 2 are different(i think), you should use different selector in them. Then expose the deployments. For example:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  labels:
    app: nginx
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.15.4
        ports:
        - containerPort: 80

--

apiVersion: apps/v1
kind: Deployment
metadata:
  name: hello-deployment
  labels:
    app: hello
spec:
  replicas: 3
  selector:
    matchLabels:
      app: hello
  template:
    metadata:
      labels:
        app: hello
    spec:
      containers:
      - name: hello
        image: nightfury1204/hello_server
        args:
        - serve
        ports:
        - containerPort: 8080

Above two deployment nginx-deployment and hello-deployment has different selector. So expose them to service will not colide each other.

When you use kubectl expose deployment app1-deployment --type=ClusterIP --name=app1-svc to expose deployment, the service will have the same selector as the deployment.

-- nightfury1204
Source: StackOverflow