This page describes Kubernetes StatefulSet objects and their use in Google Kubernetes Engine.
What is a StatefulSet?
StatefulSets represent a set of [Pods] with unique, persistent identities and stable hostnames that GKE maintains regardless of where they are scheduled. The state information and other resilient data for any given StatefulSet Pod is maintained in persistent disk storage associated with the StatefulSet.
StatefulSets use an ordinal index for the identity and ordering of their Pods.
By default, StatefulSet Pods are deployed in sequential order and are terminated
in reverse ordinal order. For example, a StatefulSet named
web has its Pods
web-2. When the
web Pod specification is
changed, its Pods are gracefully stopped and recreated in an ordered way; in
web-2 is terminated first, then
web-1, and so on.
Alternatively, you can specify the
podManagementPolicy: Parallel field
to have a StatefulSet launch or terminate all of its Pods in parallel, rather
than waiting for Pods to become Running and Ready or to be terminated prior to
launching or terminating another Pod.
StatefulSets 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.
StatefulSets are designed to deploy stateful applications and clustered applications that save data to persistent storage, such as Google Compute Engine persistent disks. StatefulSets are suitable for deploying Kafka, MySQL, Redis, ZooKeeper, and other applications needing unique, persistent identities and stable hostnames. For stateless applications, use Deployments.
Once created, the StatefulSet ensures that the desired number of Pods are running and available at all times. The StatefulSet automatically replaces Pods that fail or are evicted from their nodes, and automatically associates new Pods with the storage resources, resource requests and limits, and other configurations defined in the StatefulSet's Pod specification.
The following is an example of a Service and StatefulSet manifest file:
apiVersion: v1 kind: Service metadata: name: nginx labels: app: nginx spec: ports: - port: 80 name: web clusterIP: None selector: app: nginx --- apiVersion: apps/v1 kind: StatefulSet metadata: name: web spec: selector: matchLabels: app: nginx # Label selector that determines which Pods belong to the StatefulSet # Must match spec: template: metadata: labels serviceName: "nginx" replicas: 3 template: metadata: labels: app: nginx # Pod template's label selector spec: terminationGracePeriodSeconds: 10 containers: - name: nginx image: gcr.io/google_containers/nginx-slim:0.8 ports: - containerPort: 80 name: web volumeMounts: - name: www mountPath: /usr/share/nginx/html volumeClaimTemplates: - metadata: name: www spec: accessModes: [ "ReadWriteOnce" ] resources: requests: storage: 1Gi
In this example:
- A Service object named
nginxis created, indicated by the
metadata: namefield. The Service targets an app called
nginx, indicated by
labels: app: nginxand
selector: app: nginx. The Service exposes port 80 and names it
web. This Service control the network domain and to route Internet traffic to the containerized application deployed by the StatefulSet.
- A StatefulSet named
webis created with three replicated Pods (
- The Pod template (
spec: template) indicates that its Pods are labelled
- The Pod specification (
template: spec) indicates that the StatefulSet's Pods run one container,
nginx, which runs the
nginx-slimimage at version
0.8. The container image is hosted by Container Registry.
- The Pod specification uses the
webport opened by the Service.
template: spec: volumeMountsspecifies a
mountPath, which is named
mountPathis the path within the container at which a storage volume should be mounted.
- The StatefulSet provisions a PersistentVolumeClaim,
www, with 1GB of provisioned storage.
In sum, the Pod specification contains the following instructions:
- Label each Pod as
- In each Pod, run one container named
- Run the
nginx-slimimage at version
- Have Pods use port
- Save data to the mount path.
For more information about StatefulSet configurations, refer to the StatefulSet API reference.
You can update a StatefulSet by making changes to its Pod specification, which
includes its container images and volumes. You can also update the object's
resource requests and limits, labels, and annotations. To update a StatefulSet,
you can use
kubectl, the Kubernetes API, or the
GKE Workloads menu in Google Cloud Platform Console.
To decide how to handle updates, StatefulSets use a update strategy defined in
spec: updateStrategy. There are two strategies,
OnDeletedoes not automatically delete and recreate Pods when the object's configuration is changed. Instead, you must manually delete the old Pods to cause the controller to create updated Pods.
RollingUpdateautomatically deletes and recreates Pods when the object's configuration is changed. New Pods must be in Running and Ready states before their predecessors are deleted. With this strategy, changing the Pod specification automatically triggers a rollout. This is the default update strategy for StatefulSets.
StatefulSets update Pods in reverse ordinal order. You can monitor update rollouts by running the following command:
kubectl rollout status statefulset [STATEFULSET_NAME]
Partitioning rolling updates
You can partition rolling updates. Partitioning is useful if you want to stage an update, roll out a canary, or perform a phased roll out.
When you partition an update, all Pods with an ordinal greater than or equal to the partition value are updated when you update the StatefulSet’s Pod specification. Pods with an ordinal less than the partition value are not updated and, even if they are deleted, are recreated using the previous version of the specification. If the partition value is greater than the number of replicas, the updates are not propagated to the Pods.