Networking

Kubernetes is networking addresses four concerns

Containers within a pod use networking (internal network) to communicate via loopback
Cluster networking provides communication between different pods
The service resources let you expose an application running in pods to be reachable outside your cluster
You can also use services to publish services only for consumption inside your cluster

Communication between containers within the same pod

It usually happen through localhost within the containers.

k8s_example_10.yml

kind: Pod
apiVersion: v1
metadata:
  name: testpod
spec:
  containers:
    - name: c00
      image: ubuntu
      command: ["/bin/bash", "-c", "while true; do echo Hello-from-ubuntu; sleep 5; done"]
    - name: c01
      image: httpd
      ports:
        - containerPort: 80

Create pod

kubectl apply -f k8s_example_10.yml

Output

controlplane $ kubectl apply -f  k8s_example_10.yml
pod/example-10 created

Verify if the httpd service is running using ubuntu container

kubectl exec -it example-10 -c c00 -- /bin/bash

Here -c is to chose the specific container

Output {8}

controlplane $ kubectl exec -it example-10 -c c00 -- /bin/bash
root@example-10:/# apt update && apt install curl -y
.
.
.
done.
root@example-10:/# curl localhost:80
<html><body><h1>It works!</h1></body></html>

Communication between containers in different pod within same node

It usually happen through IP Address and the exposed port of container.
By default pods IP will not be accessible outside the node
For example - curl IP_ADDRESS:PORT

k8s_example_11_C1.yml

kind: Pod
apiVersion: v1
metadata:
  name: example-11-c1
spec:
  containers:
    - name: c00
      image: ubuntu
      command: ["/bin/bash", "-c", "while true; do echo Hello-from-ubuntu; sleep 5; done"]

k8s_example_11_C2.yml

kind: Pod
apiVersion: v1
metadata:
  name: example-11-c2
spec:
  containers:
    - name: c01
      image: httpd
      ports:
        - containerPort: 80

Create pod

kubectl apply -f k8s_example_11_C1.yml
kubectl apply -f k8s_example_11_C2.yml

Output

controlplane $ kubectl apply -f  k8s_example_11_C1.yml
pod/example-11-c1 created
controlplane $ kubectl apply -f  k8s_example_11_C2.yml
pod/example-11-c2 created

Retrieve the pods IP Address

kubectl get pods -o wide

Verify if the httpd service is running using ubuntu container

kubectl exec -it example-11 -- /bin/bash

Output

controlplane $ kubectl get pods -o wide
NAME            READY   STATUS    RESTARTS   AGE     IP            NODE     NOMINATED NODE   READINESS GATES
example-11-c1   1/1     Running   0          39s     192.168.1.8   node01   <none>           <none>
example-11-c2   1/1     Running   0          25s     192.168.1.9   node01   <none>           <none>

kubectl exec -it example-11-c1 -- /bin/bash

Here -c is to chose the specific container

Output

controlplane $ kubectl exec -it example-11-c1 -- /bin/bash
root@example-10:/# apt update && apt install curl -y
.
.
.
done.
root@example-10:/# curl 192.168.1.9:80
<html><body><h1>It works!</h1></body></html>

Important note

There is a possibility of pod crash so that the ip address will not be consistent and it will be changed dynamically from time to time in the case of pod failure. In order to make it accessible all the time via application without worrying about the change in pod IP,

We use an object called services to make the IP consistent.

Deployment Object Services Object