DaemonSet

What is a DaemonSet?

Like other workload objects, DaemonSets manage groups of replicated Pods. However, DaemonSets attempt to adhere to a one-Pod-per-node model, either across the entire cluster or a subset of nodes. As you add nodes to a node pool, DaemonSets automatically add Pods to the new nodes as needed.

DaemonSets use a Pod template, which contains a specification for its Pods. The Pod specification determines how each Pod should look: what applications should run inside its containers, which volumes it should mount, its labels and selectors, and more.

Usage patterns

DaemonSets are useful for deploying ongoing background tasks that you need to run on all or certain nodes, and which do not require user intervention. Examples of such tasks include storage daemons like ceph, log collection daemons like fluentd, and node monitoring daemons like collectd.

For example, you could have DaemonSets for each type of daemon run on all of your nodes. Alternatively, you could run multiple DaemonSets for a single type of daemon, but have them use different configurations for different hardware types and resource needs.

Creating DaemonSets

cat << EOF > daemonset.yaml
apiVersion: extensions/v1beta1 # For Kubernetes version 1.9 and later, use apps/v1
kind: DaemonSet
metadata:
  name: fluentd
spec:
  selector:
      matchLabels:
        name: fluentd # Label selector that determines which Pods belong to the DaemonSet
  template:
    metadata:
      labels:
        name: fluentd # Pod template's label selector
    spec:
      nodeSelector:
        type: prod # Node label selector that determines on which nodes Pod should be scheduled
                   # In this case, Pods are only scheduled to nodes bearing the label "type: prod"
      containers:
      - name: fluentd
        image: gcr.io/google-containers/fluentd-elasticsearch:1.20
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
EOF

kubectl create -f daemonset.yaml
daemonset.extensions "fluentd" created

In sum, the Pod specification contains the following instructions:

• Label Pod as fluentd.
• Use node label selector type: prod to schedule Pod to matching nodes, and do not schedule on nodes which do not bear the label selector. (Alternatively, omit the nodeSelector field to schedule on all nodes.)
• Run fluentd-elasticsearch at version 1.20.
• Request some memory and CPU resources.

kubectl get ds
NAME      DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
fluentd   0         0         0         0            0           type=prod       47m

kubectl get nodes
NAME                                                  STATUS    ROLES     AGE       VERSION
gke-persistent-disk-tuto-default-pool-e2669282-4g5m   Ready     <none>    1h        v1.8.10-gke.0
gke-persistent-disk-tuto-default-pool-e2669282-75xl   Ready     <none>    1h        v1.8.10-gke.0
gke-persistent-disk-tuto-default-pool-e2669282-knp6   Ready     <none>    1h        v1.8.10-gke.0

kubectl label node gke-persistent-disk-tuto-default-pool-e2669282-4g5m type=prod
node "gke-persistent-disk-tuto-default-pool-e2669282-4g5m" labeled

kubectl get ds
NAME      DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
fluentd   1         1         1         1            1           type=prod       49m

Updating DaemonSets (TODO)

You can update DaemonSets by changing its Pod specification, resource requests and limits, labels, and annotations.

DaemonSet has two update strategy types:

• OnDelete
• RollingUpdate

OnDelete

This is the default update strategy for backward-compatibility. With OnDelete update strategy, after you update a DaemonSet template, new DaemonSet pods will only be created when you manually delete old DaemonSet pods. This is the same behavior of DaemonSet in Kubernetes version 1.5 or before.

Find the update strategy of fluentd ds

kubectl get ds/fluentd -o go-template='{{.spec.updateStrategy.type}}{{"\n"}}'
OnDelete

RollingUpdate

With RollingUpdate update strategy, after you update a DaemonSet template, old DaemonSet pods will be killed, and new DaemonSet pods will be created automatically, in a controlled fashion.

cat << EOF > rolling_ds.yaml
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
  name: frontend
spec:
  updateStrategy: RollingUpdate
    maxUnavailable: 1
    minReadySeconds: 0
  template:
    metadata:
      labels:
        app: frontend-webserver
    spec:
      nodeSelector:
        app: frontend-node
      containers:
        - name: webserver
          image: nginx
          ports:
          - containerPort: 80
EOF

kubectl create -f rolling_ds.yaml
error: error converting YAML to JSON: yaml: line 6: mapping values are not allowed in this context