Rolling update

Q: How deeply does Acing the Certified Kubernetes Administrator Exam, Second Edition cover rolling update?

Works through image changes, ReplicaSet transitions, revision history, rollback, and status commands; illustrates rollout limits without exploring controller internals.

As explained in Acing the Certified Kubernetes Administrator Exam, Second Edition

An application rollout is the controlled, versioned progression of deploying a new workload or application version through successive revisions.

A rolling update replaces the pods of a Deployment as a changed application version is rolled out. It helps minimize downtime: the Deployment creates new pods for the updated version rather than requiring an administrator to replace every pod manually. This is an operational task for keeping a running application current while maintaining availability. section 5.1.1 section 5.0

A Deployment manages pods through ReplicaSets. Each ReplicaSet maintains the desired number of pods for a particular version. Changing a Deployment’s image creates a new ReplicaSet and new pods; the old ReplicaSet can remain visible with zero desired pods after the transition. A RollingUpdate strategy can set max surge and max unavailable values to control how many replicas may be added or unavailable during the update. section 5.1.2

Update an image and verify the rollout

Run `kubectl set image deploy apache httpd=httpd:alpine` to change the `httpd` container’s image. Then run `kubectl get po` to see the new pods and `kubectl get rs` to compare the new ReplicaSet with the old one. Check the Deployment’s `READY`, `UP-TO-DATE`, and `AVAILABLE` columns with `kubectl get deploy`; use `kubectl rollout status deploy apache` to check whether the rollout finishes. section 5.1.1 section 5.1.2 section 5.1.3

root@kind-control-plane:/# kubectl get rs
NAME                DESIRED   CURRENT   READY   AGE
apache-698b8cccbd   5         5         5       28m
apache-74f79bcc68   0         0         0       45m

Compare the ReplicaSets after an update: the new one has five ready pods, while the earlier one has zero desired pods. section 5.1.2

Use revision history when a change needs explanation or recovery:

Run `kubectl rollout history deploy apache` to list recorded revisions. Add a change note with `kubectl annotate deploy apache kubernetes.io/change-cause="updated image tag"`. section 5.1.3
If the new version introduces a bug, run `kubectl rollout undo deploy apache` to revert. The rollback creates a new revision; revision numbers do not move backward. section 5.1.3
For an update that must wait, use `kubectl rollout pause deploy apache`, then `kubectl rollout resume deploy apache` and check its status. section 5.1.3

How Acing the Certified Kubernetes Administrator Exam, Second Edition covers it

Cover of Acing the Certified Kubernetes Administrator Exam, Second Edition

liveBook manning.com

Read Acing the Certified Kubernetes Administrator Exam, Second Edition on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Kubernetes in Action, Second EditionThe RollingUpdate strategy is a deployment update strategy that replaces application pods gradually while applying an updated pod template.
liveBook manning.com

Amazon Web Services in Action, Third Edition: An in-depth guide to AWSA rolling update replaces instances or software versions incrementally while keeping enough of the existing deployment available to avoid service interruption.
liveBook manning.com

Machine Learning Platform EngineeringA rolling update is a gradual replacement of running application instances with instances of a newer version while maintaining service availability.
liveBook manning.com

Kubernetes for DevelopersA rolling update incrementally replaces instances of an application with a new version while retaining some existing instances during the transition.
liveBook manning.com

Infrastructure as Code, Patterns and Practices: With examples in Python and TerraformA change technique that updates similar infrastructure resources incrementally, validating each resource before proceeding to the next.
liveBook manning.com

Platform Engineering on KubernetesA rolling update replaces running application instances incrementally by starting updated instances and terminating older ones after the replacements are ready.
liveBook manning.com

Bootstrapping Microservices, Second Edition: With Docker, Kubernetes, GitHub Actions, and TerraformA rolling update is a deployment technique that incrementally replaces running application instances with a new version while keeping other instances available.

Compare how 8 books explain rolling update →

Related concepts

Recreate strategycontrasts with · A recreate strategy replaces an application version by stopping all existing instances before starting instances of the new version.
consumer group rebalancingproduces · Consumer group rebalancing is the redistribution of topic-partition assignments among consumer group members after membership or partition availability changes.
blue-green deploymentincludes · Blue-green deployment is a release pattern that maintains separate active and replacement environments and switches traffic between them after the replacement is validated.
Kubernetes Deploymentused by · A Kubernetes Deployment is a declarative API resource that manages the desired state, scaling, and updates of a replicated application workload.
maxUnavailableconfigured by · maxUnavailable is a Kubernetes rolling-update setting that limits the number of Pods allowed to be unavailable during an update.
maxSurgeconfigured by · maxSurge is a Kubernetes rolling-update setting that limits the number of additional Pods allowed above the desired replica count.

This explainer was generated by Manning's knowledge graph from the text of Acing the Certified Kubernetes Administrator Exam, Second Edition. The section links point to the parts of the book it draws on.