If a Kubernetes DaemonSet is created, the defined Pod will be deployed in every single node. It is guaranteed that the running containers occupy equal resources in each node. In this scenario, the container usually works as the daemon process.

For example, the following template has an Ubuntu image container that keeps checking its memory usage half a minute at a time:

1. To build it as a DaemonSet, execute the following code block:

$ cat daemonset-free.yaml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: ram-check
spec:
  selector:
    matchLabels:
      name: checkRam
  template:
    metadata:
      labels:
        name: checkRam
    spec:
      containers:
      - name: ubuntu-free
        image: ubuntu
        command: ["/bin/bash","-c","while true; do free; sleep 30; done"]
      restartPolicy: Always

As the Job, the selector could be ignored, but it takes the values of the labels. We will always configure the restart policy of the DaemonSet as Always, which makes sure that every node has a Pod running.

2. The abbreviation of the daemonset is ds in kubectl command, use this shorter one in the CLI for convenience:

$ kubectl create -f daemonset-free.yaml
daemonset.apps "ram-check" created

$ kubectl get ds
NAME        DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
ram-check   2         2         2         2            2           <none>          5m

3. Here, we have two Pods running in separated nodes. They can still be recognized in the channel of the pod:

$ kubectl get pods -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP               NODE
ram-check-6ldng   1/1       Running   0          9m        10.233.102.130   node1
ram-check-ddpdb   1/1       Running   0          9m        10.233.75.4      node2

4. It is good for you to evaluate the result using the subcommand kubectl logs:

$ kubectl logs ram-check-6ldng
              total        used        free      shared  buff/cache   available
Mem:        3623848      790144      329076        9128     2504628     2416976
Swap:             0           0           0
              total        used        free      shared  buff/cache   available
Mem:        3623848      786304      328028        9160     2509516     2420524
Swap:             0           0           0
              total        used        free      shared  buff/cache   available
Mem:        3623848      786344      323332        9160     2514172     2415944
Swap:             0           0           0
.
.

Whenever, you add a Kubernetes node onto your existing cluster, DaemonSets will recognize and deploy a Pod automatically.

5. Let's check again current status of DaemonSets, there are two Pods that have been deployed due to having two nodes as follows:

$ kubectl get ds
NAME        DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
ram-check   2         2         2         2            2           <none>          14m

$ kubectl get nodes
NAME          STATUS    ROLES         AGE       VERSION
node1         Ready     master,node   6h        v1.10.2
node2         Ready     node          6h        v1.10.2

6. So, now we are adding one more node onto the cluster through either kubespray or kubeadm, based on your setup:

$ kubectl get nodes
NAME          STATUS    ROLES         AGE       VERSION
node1         Ready     master,node   6h        v1.10.2
node2         Ready     node          6h        v1.10.2
node3         Ready     node          3m        v1.10.2

7. A few moments later, without any operation, the DaemonSet's size become 3 automatically, which aligns to the number of nodes:

$ kubectl get ds
NAME        DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
ram-check   3         3         3         3            3           <none>          18m


$ kubectl get pods -o wide
NAME              READY     STATUS    RESTARTS   AGE       IP               NODE
ram-check-6ldng   1/1       Running   0          18m       10.233.102.130   node1
ram-check-ddpdb   1/1       Running   0          18m       10.233.75.4      node2
ram-check-dpdmt   1/1       Running   0          3m        10.233.71.0      node3