As we discussed the pods and their manual deployment, I have to tell you something. Take a deep breath, ok…and sit probably :).

NO ONE WORKS WITH PODS

I will let that rest for a second….

Ok, now, what I mean. The reason, why no one works with pods is because, they are too granular. What if we need to put 5 replicas of that image on 5 workers. Well, there is something called Replication controller, which encapsulates the pod and add more features.

So let's see visual representation of what I mean:

With the replication controller, we can define N amount of replicas and the controller takes care of the rest. The controller will ensure our desired state (5 replicas) is always in place.

So how we do it ? Well the process is pretty similar to pods as it relies on them.

1. We create a manifest file
2. We present it to the API
3. The API do the rest

As with the manual configuration of the pods, we have to prepare a little manifest file. I have prepared a simple one:

apiVersion: v1
kind: ReplicationController
metadata:
  name: hello-rc
spec:
  replicas: 5
  selector:
    app: hello-date
  template:
    metadata:
      labels:
        app: hello-date
    spec:
      containers:
      - name: hello-pod
        image: andonovj/httpserverdemo:latest
        ports:
        - containerPort: 1234

The first information is about the replication controller. In our case:

1. it is Version 1
2. it is Replication Controller (you to believe :) )
3. It is called: “hello-rc”
4. It has 5 replicas, so it will configure 5 pods
5. It will run application called: “hello-date” defined in the template

Later we have a pod a template section which has:

• Metadata
• Specification

From here, you can orient I hope :)

You see that the replication controller gives us way more possibilities and furthermore it relies on the Pods as it is using them.

The command to create the controller, once you have the YML/JSON file is pretty similar as the pod:

ubuntu@k8s-master:~$ kubectl create -f rc.yml
replicationcontroller/hello-rc created
ubuntu@k8s-master:~$

After that, we can check the status and the pods

ubuntu@k8s-master:~$ kubectl get rc
NAME       DESIRED   CURRENT   READY   AGE
hello-rc   5         5         3       17s
ubuntu@k8s-master:~$ kubectl describe rc
Name:         hello-rc
Namespace:    default
Selector:     app=hello-date
Labels:       app=hello-date
Annotations:  <none>
Replicas:     5 current / 5 desired
Pods Status:  5 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  app=hello-date
  Containers:
   hello-pod:
    Image:        andonovj/httpserverdemo:latest
    Port:         8080/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Events:
  Type    Reason            Age   From                    Message
  ----    ------            ----  ----                    -------
  Normal  SuccessfulCreate  10m   replication-controller  Created pod: hello-rc-dlx7j
  Normal  SuccessfulCreate  10m   replication-controller  Created pod: hello-rc-2kqmm
  Normal  SuccessfulCreate  10m   replication-controller  Created pod: hello-rc-twz2w
  Normal  SuccessfulCreate  10m   replication-controller  Created pod: hello-rc-8rn9t
  Normal  SuccessfulCreate  10m   replication-controller  Created pod: hello-rc-j2swd
ubuntu@k8s-master:~$

In our case, we have 5 pods on our 2 workers:

• 2 on Worker 1
• 3 on Worker 2

We can update the configuration, by simply updating the YML/JSON file. For example, let's increase the replicas from 5 → 10, firstly we update it in the YML/JSON file:

spec:
  replicas: 10

After that, we simply apply that change:

ubuntu@k8s-master:~$ kubectl apply -f rc.yml
Warning: kubectl apply should be used on resource created by either kubectl create --save-config or kubectl apply
replicationcontroller/hello-rc configured
ubuntu@k8s-master:~$

You can ignore the warning line :) You can check the configuration as follows:

ubuntu@k8s-master:~$ kubectl get rc
NAME       DESIRED   CURRENT   READY   AGE
hello-rc   10        10        10      16m
ubuntu@k8s-master:~$ kubectl describe rc
Name:         hello-rc
Namespace:    default
Selector:     app=hello-date
Labels:       app=hello-date
Annotations:  Replicas:  10 current / 10 desired
Pods Status:  10 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  app=hello-date
  Containers:
   hello-pod:
    Image:        andonovj/httpserverdemo:latest
    Port:         1234/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Events:
  Type    Reason            Age   From                    Message
  ----    ------            ----  ----                    -------
  Normal  SuccessfulCreate  16m   replication-controller  Created pod: hello-rc-dlx7j
  Normal  SuccessfulCreate  16m   replication-controller  Created pod: hello-rc-2kqmm
  Normal  SuccessfulCreate  16m   replication-controller  Created pod: hello-rc-twz2w
  Normal  SuccessfulCreate  16m   replication-controller  Created pod: hello-rc-8rn9t
  Normal  SuccessfulCreate  16m   replication-controller  Created pod: hello-rc-j2swd
  Normal  SuccessfulCreate  73s   replication-controller  Created pod: hello-rc-zhngb
  Normal  SuccessfulCreate  73s   replication-controller  Created pod: hello-rc-vxzfl
  Normal  SuccessfulCreate  73s   replication-controller  Created pod: hello-rc-dklgj
  Normal  SuccessfulCreate  73s   replication-controller  Created pod: hello-rc-cf8jw
  Normal  SuccessfulCreate  72s   replication-controller  Created pod: hello-rc-6tg65
ubuntu@k8s-master:~$

You see that some pods are newer than other, these are the newly created ones :)

TODO