Certified Kubernetes Administrator - ReplicationControllers, ReplicaSets, Deployments


Kubernetes was designed in a highly modular manner to orchestrate multiple containers for a specific purpose. In this post, I'll focus on ReplicationController, ReplicaSets, and deployments.

Kubernetes needs replication for;

Before we go in explain how to create and manage replicas let's try to explain why we need to create multiple replicated containers. Replication is used for the core purpose of;

  • Reliability and Availability: When you create multiple replicated pods, you probably minimize the effects of one or more pods that fails on a host. 
  • Load Balancing: Running multiple pods on different nodes enables you to increase the capacity and reliability of applications. Balancing methods will prevent overloading of a single node and enables you to perform maintenance task without downtime.
  • Scaling and Maintaining: Scaling becomes an important part of the application. Kubernetes enables you to easily create scale up/down scenarios that application can handle more and more requests per minute.

Kubernetes Replication

Kubernetes has multiple ways in which you can implement replication to the container base application. In this post, we will explain ReplicationControllers, ReplicaSets, and Deployments. 

ReplicationController

A ReplicationControllers ensures that a specified number of pod replicas are running at any one time. In other words, a ReplicationController makes sure that a pod or a homogeneous set of pods is always up and available. You can get more detail about how a ReplicationController works from Kubernetes Documentation.

As with all other Kubernetes definition files, a ReplicationController needs apiVersion, kind, metadata and specs fields. A ReplicationController has two spec fields. The first spec fields contain ReplicationControllers information like selector, template, replicas, etc. The second one will be used to define container base information like image, name, ports, nodeselector, storage, environment, etc. 

apiVersion: v1
kind: ReplicationController
metadata:
  name: nginx
# modify replicas according to your case
spec:
  replicas: 3
  selector:
    app: nginx
  template:
    metadata:
      name: nginx
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx
        ports:
        - containerPort: 80

In this definition file, we are looking for one selector;

  • app labels must be Nginx

A simple example shows how to use kubectl command to create ReplicationControllers template file. Then you can modify this file as you need.

#kubectl  run  --generator=run/v1  --image=nginx  web-server --port=80  --dry-run  -o yaml >  /tmp/web.yaml
-kubectl run --generator=run/v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
apiVersion: v1
kind: ReplicationController
metadata:
  creationTimestamp: null
  labels:
    run: web-server
  name: web-server
spec:
  replicas: 1
  selector:
    run: web-server
  template:
    metadata:
      creationTimestamp: null
      labels:
        run: web-server
    spec:
      containers:
      - image: nginx
        name: web-server
        ports:
        - containerPort: 80
        resources: {}
status:
  replicas: 0
#kubectl  create  -f  /tmp/web.yaml
replicationcontroller/web-server created
#kubectl  get replicationcontroller
NAME         DESIRED   CURRENT   READY   AGE
web-server   1         1         1       8s
#kubectl  get  pods
NAME               READY   STATUS    RESTARTS   AGE
web-server-9rllk   1/1     Running   0          12s

To get the status of running ReplicationController use "describe" option of kubectl. As you can see there is only "1" pod running on the system named "web-server-9rllk". 

#kubectl  describe  replicationcontroller web
Name:         web-server
Namespace:    default
Selector:     run=web-server
Labels:       run=web-server
Annotations:  <none>
Replicas:     1 current / 1 desired
Pods Status:  1 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  run=web-server
  Containers:
   web-server:
    Image:        nginx
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Events:
  Type    Reason            Age    From                    Message
  ----    ------            ----   ----                    -------
  Normal  SuccessfulCreate  2m17s  replication-controller  Created pod: web-server-9rllk

To change replicas count use "edit"  option of kubectl.

#kubectl  edit replicationcontroller  web-server
replicationcontroller/web-server edited
#kubectl  get  pods
NAME               READY   STATUS              RESTARTS   AGE
web-server-95pfd   0/1     ContainerCreating   0          2s
web-server-9rllk   1/1     Running             0          6m19s
#kubectl get pods --selector=run=web-server --output=jsonpath={.items..metadata.name}
web-server-95pfd web-server-9rllk

 

ReplicaSets

ReplicaSets and ReplicationControllers are used for the core purpose of reliability, availability, load balancing, scaling and maintaining. ReplicaSet is the next-generation ReplicationController that supports the new set-based label selector. Basically ReplicaSets has more options on the selector field when compare ReplicationControllers.

Kubernetes recommends using Deployments instead of directly using ReplicaSets or ReplicationController.

apiVersion: apps/v1
kind: ReplicaSet
metadata:
  name: nginx
spec:
  replicas: 3
  selector:
    matchExpressions:
       - {key: app, operator: In, values: [nginx]}
       - {key: environment, operator: NotIn, values: [production]}
  template:
    metadata:
      name: nginx
      labels:
        app: nginx
        environment: production
    spec:
      containers:
      - name: nginx
        image: nginx
        ports:
        - containerPort: 80

In this definition file, we are looking for two different selectors;

  • app labels must be Nginx
  • environment label must not be production

A simple example shows how to use kubectl command to create ReplicaSets template file. Use same command to create a  template file. That change it as you need.

#kubectl  run --generator=run/v1  --image=nginx  web-server-replicaset --port=80  --dry-run  -o yaml >  /tmp/web_server_rs.yaml
#cat /tmp/web_server_rs.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
  creationTimestamp: null
  labels:
    app: web-server-replicaset
  name: web-server-replicaset
spec:
  replicas: 1
  selector:
    matchExpressions:
       - {key: app, operator: In, values: [web-server-replicaset]}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: web-server-replicaset
        environment: production
    spec:
      containers:
      - image: nginx
        name: web-server-replicaset
        ports:
        - containerPort: 80
        resources: {}
status:
  replicas: 0
#kubectl  create  -f  /tmp/web_server_rs.yaml
replicaset.apps/web-server-replicaset created
#kubectl  get  rs
NAME                    DESIRED   CURRENT   READY   AGE
web-server-replicaset   1         1         1       13s
#kubectl  get  pods
NAME                          READY   STATUS    RESTARTS   AGE
web-server-replicaset-ggwjr   1/1     Running   0          20s
#kubectl  describe  rs/web-server-replicaset
Name:         web-server-replicaset
Namespace:    default
Selector:     app in (web-server-replicaset)
Labels:       app=web-server-replicaset
Annotations:  <none>
Replicas:     1 current / 1 desired
Pods Status:  1 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  app=web-server-replicaset
           environment=production
  Containers:
   web-server-replicaset:
    Image:        nginx
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Events:
  Type    Reason            Age   From                   Message
  ----    ------            ----  ----                   -------
  Normal  SuccessfulCreate  84s   replicaset-controller  Created pod: web-server-replicaset-ggwjr
#kubectl edit  rs/web-server-replicaset
replicaset.apps/web-server-replicaset edited
#kubectl  get  pods
NAME                          READY   STATUS              RESTARTS   AGE
web-server-replicaset-ggwjr   1/1     Running             0          2m30s
web-server-replicaset-vjwh6   0/1     ContainerCreating   0          1s
#kubectl delete  rs/web-server-replicaset
replicaset.apps "web-server-replicaset" deleted

 

Deployments

This method is the easiest and most used way to create resource for deploying applications. Deployments provide the same replication function like ReplicaSets but also additional ability like rollout and rollback if necessary. For example, if you create a deployment with one replica, it will check the desire and the current state of ReplicaSet. So if the current state is "0" then it will create ReplicaSet and  Pod. Deployment creates a ReplicaSet with name  <Deployment_Name>-<ReplicaSetID>. Also, Replicaset will create pods named <Deployment_Name>-<ReplicaSetID>-<PodID>.

We will use the following definition file to create a basic deployment. You can use it by copying the added definition file and save it as nginx_deployment.yaml. Use Katacoda to exercise.

#cat > nginx_deployment.yaml <<EOF
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:1.7.9
        ports:
        - containerPort: 80
EOF
#kubectl  create  -f  nginx_deployment.yaml
deployment.apps/nginx-deployment created
#kubectl  get deployment
NAME               READY   UP-TO-DATE   AVAILABLE   AGE
nginx-deployment   0/3     3            0           6s
#kubectl  get replicaset
NAME                          DESIRED   CURRENT   READY   AGE
nginx-deployment-54f57cf6bf   3         3         3       13s
#kubectl  get pods
NAME                                READY   STATUS    RESTARTS   AGE
nginx-deployment-54f57cf6bf-6k7f7   1/1     Running   0          20s
nginx-deployment-54f57cf6bf-h6fbf   1/1     Running   0          20s
nginx-deployment-54f57cf6bf-jhnl5   1/1     Running   0          20s

Let's go ahead and dig new attributes of deployment. 

  • StrategyType: Specifies the strategy used to replace old Pods by new ones. RollingUpdate or  Recreate can be defined. Check Kubernetes Documentation for more details. 
  • RollingUpdateStrategy: You can specify maxUnavailable and maxSurge to control the rolling update process
  • MinReadySeconds: The minimum number of seconds for which a newly created Pod should be ready without any of its containers crashing, for it to be considered available. The default value is "0".

 

#kubectl  describe  deployment nginx-deployment
Name:                   nginx-deployment
Namespace:              default
CreationTimestamp:      Sun, 12 Jan 2020 19:17:11 +0000
Labels:                 app=nginx
Annotations:            deployment.kubernetes.io/revision: 1
Selector:               app=nginx
Replicas:               3 desired | 3 updated | 3 total | 3 available | 0 unavailable
StrategyType:           RollingUpdate
MinReadySeconds:        0
RollingUpdateStrategy:  25% max unavailable, 25% max surge
Pod Template:
  Labels:  app=nginx
  Containers:
   nginx:
    Image:        nginx:1.7.9
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:       <none>
  Volumes:        <none>
Conditions:
  Type           Status  Reason
  ----           ------  ------
  Available      True    MinimumReplicasAvailable
  Progressing    True    NewReplicaSetAvailable
OldReplicaSets:  <none>
NewReplicaSet:   nginx-deployment-54f57cf6bf (3/3 replicas created)
Events:
  Type    Reason             Age    From                   Message
  ----    ------             ----   ----                   -------
  Normal  ScalingReplicaSet  2m10s  deployment-controller  Scaled up replica set nginx-deployment-54f57cf6bf to 3

 

Time to test your skills. Use  Katacoda and find out the answers.

 

Question 1:  Create a namespace called casesup?

#kubectl  create namespace  casesup
namespace/casesup created
#kubectl describe  namespace casesup
Name:         casesup
Labels:       
Annotations:  
Status:       Active

No resource quota.

No LimitRange resource.

Question 2: Use casesup namespcae and create a replicationcontroller. You can use busybox image to create pod. Also get the pods IP and log it.

#kubectl  run  --generator=run/v1  --image=busybox busy-replicationcontroller --dry-run -o yaml >  /tmp/replicationcontroller.yaml
#cat /tmp/replicationcontroller.yaml
apiVersion: v1
kind: ReplicationController
metadata:
  namespace: casesup
  creationTimestamp: null
  labels:
    run: busy-replicationcontroller
  name: busy-replicationcontroller
spec:
  replicas: 1
  selector:
    run: busy-replicationcontroller
  template:
    metadata:
      creationTimestamp: null
      labels:
        run: busy-replicationcontroller
    spec:
      containers:
      - image: busybox
        command: ["/bin/sh"]
        args: ["-c", "while true; do echo $(ip a|grep inet|grep global|awk  '{print $2}');sleep 2; done"]
        name: busy-replicationcontroller
        resources: {}
status:
  replicas: 0
#kubectl  create  -f /tmp/replicationcontroller.yaml
#kubectl  get replicationcontroller  -n casesup
NAME                         DESIRED   CURRENT   READY   AGE
busy-replicationcontroller   1         1         1       4m22s
#kubectl  get  pods -n casesup
NAME                               READY   STATUS    RESTARTS   AGE
busy-replicationcontroller-x2pb6   1/1     Running   0          4m33s
#kubectl  logs  busy-replicationcontroller-x2pb6 -n casesup
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24
172.18.0.7/24 

-Use "edit" option to modify ReplicationControllers 
#kubectl edit  replicationcontroller busy-replicationcontroller

Question 3:  Use casesup namespcae and create a ReplicaSets. You can use busybox image to create pod. ReplicaSets selector options will be "app=busybox environment=test"

#kubectl  run  --generator=run/v1  --image=busybox busy-replicationcontroller --dry-run -o yaml >  /tmp/replicasets.yaml
kubectl run --generator=run/v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
-Change some  parameters 
#vi  /tmp/replicasets.yaml
#cat /tmp/replicasets.yaml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
  namespace: casesup
  labels:
    app: busybox
  name: busy-replicasets
spec:
  replicas: 2
  selector:
    matchExpressions:
        - {key: app, operator: In, values: [busybox]}
        - {key: environment, operator: In, values: [test]}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: busybox
        environment: test
    spec:
      containers:
      - image: busybox
        command: ["/bin/sh"]
        args: ["-c", "while true; do echo $(ip a|grep inet|grep global|awk  '{print $2}');sleep 2; done"]
        name: busy-replicasets
        resources: {}
status:
  replicas: 0

#kubectl  create  -f  /tmp/replicasets.yaml
replicaset.apps/busy-replicasets created

#kubectl  get replicasets  -n casesup
NAME               DESIRED   CURRENT   READY   AGE
busy-replicasets   2         2         2       6s

#kubectl  get pods  -n casesup
NAME                     READY   STATUS    RESTARTS   AGE
busy-replicasets-q8rs4   1/1     Running   0          11s
busy-replicasets-qjvk6   1/1     Running   0          11s

Question 4:  Use busybox-deployment that you created before. Change container port and add "change-cause" to the deployment config. Use "kubectl get deployment" command to figure out how pods will terminates. Perform a rollback operation for the busybox-deployment.

--You can add "kubernetes.io/change-cause: port changed  from 8080  to 80"  to the metadata of deployment definition file. Also change  port from 80 to 8080 from container port.
#kubectl edit  deployment  busybox-deployment  -n casesup
deployment.extensions/busybox-deployment edited
#kubectl  get pods -n casesup
NAME                                  READY   STATUS              RESTARTS   AGE
busybox-deployment-6c8769f9df-kpwfk   1/1     Running             0          6m43s
busybox-deployment-6c8769f9df-ltgl2   1/1     Running             0          6m49s
busybox-deployment-6c8769f9df-nmw6x   1/1     Running             0          6m46s
busybox-deployment-7f656ff6bf-b6c4w   0/1     ContainerCreating   0          2s

#kubectl  get pods -n casesup
NAME                                  READY   STATUS              RESTARTS   AGE
busybox-deployment-6c8769f9df-kpwfk   1/1     Running             0          6m44s
busybox-deployment-6c8769f9df-ltgl2   1/1     Running             0          6m50s
busybox-deployment-6c8769f9df-nmw6x   1/1     Running             0          6m47s
busybox-deployment-7f656ff6bf-b6c4w   0/1     ContainerCreating   0          3s

#kubectl  get pods -n casesup
NAME                                  READY   STATUS              RESTARTS   AGE
busybox-deployment-6c8769f9df-kpwfk   1/1     Terminating         0          6m45s
busybox-deployment-6c8769f9df-ltgl2   1/1     Running             0          6m51s
busybox-deployment-6c8769f9df-nmw6x   1/1     Running             0          6m48s
busybox-deployment-7f656ff6bf-46mwg   0/1     ContainerCreating   0          1s
busybox-deployment-7f656ff6bf-b6c4w   1/1     Running             0          4s

#kubectl  get pods -n casesup
NAME                                  READY   STATUS        RESTARTS   AGE
busybox-deployment-6c8769f9df-kpwfk   1/1     Terminating   0          6m54s
busybox-deployment-6c8769f9df-ltgl2   1/1     Terminating   0          7m
busybox-deployment-6c8769f9df-nmw6x   1/1     Terminating   0          6m57s
busybox-deployment-7f656ff6bf-46mwg   1/1     Running       0          10s
busybox-deployment-7f656ff6bf-b6c4w   1/1     Running       0          13s
busybox-deployment-7f656ff6bf-dwk7v   1/1     Running       0          7s

#kubectl rollout history deployment/busybox-deployment -n casesup
deployment.extensions/busybox-deployment
REVISION  CHANGE-CAUSE
2         port changed  from 8080  to 80
-Performed  rollback operation
#kubectl  rollout  undo  deployment  busybox-deployment -n casesup
deployment.extensions/busybox-deployment rolled back
#kubectl  get pods -n casesup
NAME                                  READY   STATUS        RESTARTS   AGE
busybox-deployment-6c8769f9df-2nbz7   1/1     Running       0          8s
busybox-deployment-7f656ff6bf-46mwg   1/1     Running       0          2m36s
busybox-deployment-7f656ff6bf-b6c4w   1/1     Running       0          2m39s
busybox-deployment-7f656ff6bf-dwk7v   1/1     Terminating   0          2m33s
#kubectl  get pods -n casesup
NAME                                  READY   STATUS              RESTARTS   AGE
busybox-deployment-6c8769f9df-2nbz7   1/1     Running             0          12s
busybox-deployment-6c8769f9df-vqtxr   0/1     ContainerCreating   0          4s
busybox-deployment-7f656ff6bf-46mwg   1/1     Running             0          2m40s
busybox-deployment-7f656ff6bf-b6c4w   1/1     Running             0          2m43s
busybox-deployment-7f656ff6bf-dwk7v   1/1     Terminating         0          2m37s

I'm a IT Infrastructure and Operations Architect with extensive experience and administration skills and works for Turk Telekom. I provide hardware and software support for the IT Infrastructure and Operations tasks.

205 Total Posts
Follow Me

Related Post

0 Comments

Leave a Reply