v0.2.5

⚠️ Breaking Changes

• Build master tag on master instead of latest (#1513): users that have manifests pointing to the latest tag must switch to release-0.2 for v0.2.x and to master if you want to follow ongoing development. Note, this is only for the bleeding edge and does not apply to released versions such as v0.2.5.

✨ New Features

• Machine controller: Drain node before machine deletion (#1096)

📖 Documentation

• Proposal v2: expectations and timelines (#1501)
• Improve machine controller book (#1499)

🏃 Others

• Update controller-runtime and golang versions (#1519)
• Add i/o timeout to retry errors to acquire kube client (#1517)
• Manager pull policy temp file to gitignore (#1500)
• Cleanup the readability of some gomega error checking (#1497)
• Unit tests for machine controller (#1493)

The image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.2.5

Thanks to all our contributors! 😊

v0.2.4

Changes since v0.2.3

⚠️ Action Required

The default leader election ID has been updated in this release. When upgrading from a previous v0.2.3 (or older) release you will need to:

1. Scale down the old controllers: kubectl -n capi-system scale deployment/controller-manager --replicas=0.
2. Deploy the updated manifests with the newer image.
3. Scale up the new controllers: kubectl -n capi-system scale deployment/controller-manager --replicas=1.

Failure to scale down the Deployment prior to updating to this version will result in multiple controllers running concurrently during the roll out of the updated controller-manager image.

🐛 Bug Fixes

• Support running alongside other Cluster API pods in the same namespace with leader election enabled (#1492)
• Do not re-queue forever if only cluster is deployed (#1466)
• Update Machine Status with Error Reason and Error Message if infra object is deleted (#1458)
• DeepCopy unstructured object in patch helper Patch method (#1467)
• DeepCopy unstructured objects in helper before patching (#1464)
• Reference to PULL_BASE_REF for image building (#1457)
• Add well-known tolerations to released CAPI manager (#1487)

📖 Documentation

• Fix link in Cluster controller doc (#1479)
• Embed AWS pre-requisites document (#1472)
• Create documentation for machine contracts between components (#1391)
• Broken links to api-conventions doc (#1462)

🏃 Others

• Add common utilities for provider-based testing (#1474)
• unused function isScalingEvent (#1463)
• NodeOutOfDisk condition obsolete in Kubernetes 1.16 (#1476)

The image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.2.4

Thanks to all our contributors! 😊

v0.2.3

Changes since v0.2.2

⚠️ Breaking Changes

• Clarify Machine.Spec.ObjectMeta use in godoc (#1454)

✨ New Features

• Tag branches or latest when building images (#1447)

🐛 Bug Fixes

• Machine deletion: try up to 10s to delete the Node, then move on (#1452)
• Fix the issue feature template (#1437)
• Patch in MachineSet adoption (#1430)
• Reapply secret OwnerRef on pivot with correct UID (#1435)
• Better check to wait for the control-plane to become available (#1426)

📖 Documentation

• Changes confusing language in the CAEP template (#1448)
• Adds a cluster controller document (#1390)
• Add CABPK abbreviation and v1alpha2-ize other abbreviations (#1432)
• Updated release docs (#1424)

🏃 Others

• Deprecation message to clusterctl commands (#1439)
• Remove for machineset concurrent delete calls (#1453)
• Fix table CSS (#1449)
• Adds chuckha to maintainers list (#1409)
• Add go mod verification in CI (#1443)
• Update [email protected] and [email protected] (#1436)
• Make getMachinesInCluster a helper function (#1417)
• Update issue templates prior to v1alpha3 mass submission. (#1434)

The image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.2.3

Thanks to all our contributors! 😊

v0.2.2

Changes since v0.2.1

⚠️ Breaking Changes

• Add OwnerReference to Kubeconfig secret (#1403)
• Remove the Secret's OwnerReferences before pivot (#1411)

✨ New Features

• Update to golangci-lint v1.18.0 (#1415)
• Add util/record package from CAPA (#1413)

🐛 Bug Fixes

• track MS with old labels after updating MD labels (#1358)
• Improve handling of patching and updates in MachineSet and MachineDeployment controllers (#1380)
• manager: use our RESTMapper (#1405)
• Updates release notes generator (#1393)

📖 Documentation

• Add embed util for mdbook / fixup prerequisites for CAPA (#1410)
• Move clusterctl to book (#1397)
• Add WIP for various docs (#1401)
• Improve quickstart for AWS components installation (#1396)
• Quickstart / add other sections (#1379)
• Mark proposals as implemented (#1394)

🏃 Others

• Update status.controlPlaneInitialized from cluster controller (#1356)
• Update cloudbuild.yaml for PULL_BASE_REF substitution (#1406)

Thanks to all our contributors! 😊

The image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.2.2

v0.2.1

🚨 This is a MAJOR release.

Cluster API has some major breaking changes, which are documented in the proposals listed below. We defined the project goals and objectives and terminology.

📖 Proposals

• Machine States & Preboot Bootstrapping
• Cluster Provider Spec and Status as CRDs

📖 Documentation

• Introduce mdbook (#1376)

✨ Notable changes

• API group has been updated to cluster.x-k8s.io
• API version has been updated to v1alpha2
• Cluster API controller manager now fully manages Clusters, Machines, MachineSets, MachineDeployments

🏃 Others

• Kubebuilder project is now v2
• Controller runtime and controller tools have been updated to v0.2.0
• Go has been updated to 1.12.9
• Switched to Go Modules and removed the vendored dependencies
• Kustomize has been updated to v3.1.0
• clusterctl now pivots additional resources, such as: cluster-related secrets, bootstrap, and infrastructure custom resources

Thanks to all our contributors! 😊

The image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.2.1

v0.1.10

Changes since v0.1.9:

• #1302: Support concurrency in cluster and machine controllers
• #1357: Pivot secrets

The container image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.1.10

v0.1.9

Fixes since v0.1.8

• #1185: Added support for kustomize v3.0.3+
• #1180: Added logic to delete a Cluster's MachineDeployments, MachineSets, and Machines as part of Cluster deletion
• #1207: Clusters will set OwnerReferences on Machines, MachineSets, and MachineDeployments if they are not already owned by another MachineSet or MachineDeployment

The controller manager image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.1.9

v0.1.8

Fixes since v0.1.7

• #1173: Get updated cluster before removing finalizers

The controller manager image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.1.8

v0.1.7

Fixes since v0.1.6

• #1154: Don't swallow errors when reconciling noderef

The controller manager image for this release is: us.gcr.io/k8s-artifacts-prod/cluster-api/cluster-api-controller:v0.1.7

v0.1.6

Fixes since v0.1.4

• Fixed issue handling provider IDs of the form :// (specifically, 2 slashes)
• Fixed issue where deleting a Machine could hang if it was the last control plane member
• Fixed issue where it could take ~10 minutes per machine when rolling out an update to a MachineDeployment
• Backported support for go modules; both dep and go modules are supported for v0.1.x. Master only supports go modules.