K
Q

Kubernetes sevice with CluserIp does not pass the request to some of it's Endpoints

2/21/2020

I'm new to Kubernetes. I have setup 3 node cluster with two workers according to here.
My configurations

kubectl version
Client Version: version.Info{Major:"1", Minor:"14", GitVersion:"v1.14.0", GitCommit:"641856db18352033a0d96dbc99153fa3b27298e5", GitTreeState:"clean", BuildDate:"2019-03-25T15:53:57Z", GoVersion:"go1.12.1", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"14", GitVersion:"v1.14.10", GitCommit:"575467a0eaf3ca1f20eb86215b3bde40a5ae617a", GitTreeState:"clean", BuildDate:"2019-12-11T12:32:32Z", GoVersion:"go1.12.12", Compiler:"gc", Platform:"linux/amd64"}
 kubeadm version
kubeadm version: &version.Info{Major:"1", Minor:"14", GitVersion:"v1.14.0", GitCommit:"641856db18352033a0d96dbc99153fa3b27298e5", GitTreeState:"clean", BuildDate:"2019-03-25T15:51:21Z", GoVersion:"go1.12.1", Compiler:"gc", Platform:"linux/amd64"}

Deployed simple python service listen to 8000 port http and reply "Hello world"

my deployment config

apiVersion: apps/v1
kind: Deployment
metadata:
  name: frontend-app
  labels:
    app: frontend-app
spec:
  replicas: 2 
  selector:
    matchLabels:
      app: frontend-app
  template:
    metadata:
      labels:
        app: frontend-app
    spec:
      containers:
      - name: pyfrontend
        image: rushantha/pyfront:1.0 
        ports:
        - containerPort: 8000

Exposed this as a service kubectl expose deploy frontend-app --port 8000

I can see it deployed and running.

kubectl describe svc frontend-app
Name:              frontend-app
Namespace:         default
Labels:            app=frontend-app
Annotations:       <none>
Selector:          app=frontend-app
Type:              ClusterIP
IP:                10.96.113.192
Port:              <unset>  8000/TCP
TargetPort:        8000/TCP
Endpoints:         172.16.1.10:8000,172.16.2.9:8000
Session Affinity:  None
Events:            <none>

when I log in to each service machine and do curl pods respond ie. curl 172.16.1.10:8000 or curl 172.16.2.9:8000

However when I try to access the pods via the ClusterIp only one pod always responds. So curl sometimes hangs, most probably the other pod does not respond. I confirmed when I tail the access logs for both pods. One pod never received any requests.

curl 10.96.113.192:8000/ ---> Hangs sometimes.

Any ideas how to troubleshoot this and fix ?

-- Sajith Silva
http
kubernetes

Similar Questions

Best portable way to connect from within a pod in a local dev kubernetes cluster to docker compose service
How to disconnect already ESTABLISHED connection in kubernetes POD
Deploy Microservice to Kubernetes with PostgreSQL via using Helm
Mesos cluster deployment on Kubernetes

1 Answer

2/24/2020

After comparing the tutorial document and the outputs configuration I discovered that the --pod-network-cidr declared in the document is different from the OP endpoints which solved the problem.

The network in the flannel configuration should match the pod network CIDR otherwise pods won`t be able to communicate with each other.

Some additional information that are worth checking:

  1. Under the CIDR Notation section there is a good explanation how this system works.

  2. I find this document about networking in kuberenetes very helpful.

-- acid_fuji
Source: StackOverflow