v2.4.0

Release notes for Cluster API Provider AWS (CAPA) v2.4.0

Documentation

Changelog since v2.3.0

Urgent Upgrade Notes

(No, really, you MUST read this before you upgrade)

• v1beta1 types are not going to be served any longer and will be removed in the next release (#4757, @vincepri)

Changes by Kind

Feature

• A new enum value, disabled, is added for the AWSCluster.spec.controlPlaneLoadBalancer.loadBalancerType field, which skips the reconciliation of the load balancer for the given cluster, useful for clusters which are consuming an externally managed Control Plane. (#4438, @prometherion)
• Add & Reconcile ROSAMachinePools.spec.version (#4764, @muraee)
• Add tests for the re-introduced managed subnet functionality in #4474 (#4637, @AndiDog)
• Added AWSMachine and AWSMachineTemplate fields to control vpc placement for individual instances (#4541, @cnmcavoy)
• Adds support for a secondary, internal API server load balancer within clusters. (#4733, @nrb)
• Allow to configure garbage collector using clusterawsadm (#4500, @Fedosin)
• Enable transit encryption on S3 buckets (#4676, @fiunchinho)
• Generate CAPI kubeconfig secret for ROSA clusters (#4742, @muraee)
• Ignition: add option to store User Data in plain text (#4700, @damdo)
• Introduce basic support for ROSAMachinePools (#4725, @muraee)
• Log full ARN in GC error messages (#4709, @AndiDog)
• Prevent users setting SpotMarketOptions and MixedInstancesPolicy at the same time (#4656, @fiunchinho)
• Reconcile ROSAControlPlane version to trigger upgrades (#4761, @muraee)
• Reconcile ROSAControlePlane status (#4759, @muraee)
• Remove ingress and egress rules from vpc default security group (#4707, @fiunchinho)
• Skip AWSCluster deletion reconciliation once CAPA finalizer is gone (#4730, @AndiDog)
• Support setting HostnameType options for subnets and machines (#4664, @MaxFedotov)
• Trigger machine pool instance refresh (node rollout) if bootstrap config reference changes (#4619, @AndiDog)
• Use go 1.21.5 (#4710, @fiunchinho)
• Bump CAPI to v1.6.x (#4739, @damdo)

Bug or Regression

• Don't error when ASG is not present while deciding whether or not to refresh ASG instances. (#4660, @fiunchinho)
• Fix logging verbosity for machine pool reconciliation (#4620, @AndiDog)
• Fix: bump rosa deps to fix int overflow on non-64bit arch (#4836, @damdo)
• Make VPC creation idempotent to avoid indefinite creation of new VPCs if storage of the ID fails (#4723, @AndiDog)
• Resolved issue where deleted control planes were still registered in the kube-apiserver load balancer, causing potential disruptions. (#4701, @tobiasgiese)
• Skip instance refresh attempt if ASG does not yet exist (#4662, @fiunchinho)
• Treat the annotation cluster.x-k8s.io/replicas-managed-by on MachinePool CR as described in CAPI documentation and do not set desired replicas on ASG when the annotation is present. (#4654, @calvix)
• Update AWSMachine providerID and instanceID earlier to minimize scenarios where tag-based searching is needed (#4670, @mjlshen)
• Various changes to the releases process. (#4813, @richardcase)
• When using Ignition, the S3 bucket was previously created in us-east-1 regardless of where the cluster was located, new S3 buckets will be created within the same region as the cluster. (#4732, @vincepri)

Failing Test

• Add CI script to test docker builds for the controller and clusterawsadm with all architectures. (#4834, @richardcase)

Support

• Disable dependency report short term. (#4825, @richardcase)
• Move @AndiDog to reviewer status (#4687, @richardcase)
• Update metadata for v2.4.x release series. (#4833, @richardcase)

Other (Cleanup or Flake)

• Dependabot configuration update to group updates, (#4688, @richardcase)
• Fix some logging and error cases (#4724, @AndiDog)
• If you're using S3 buckets for bootstrap data, the following permissions are required on the controller to automatically detect whether an object is still needing to be deleted: s3:GetObject. The following permissions are optional: s3:ListBucket. (#4667, @thefirstofthe300)

Uncategorized

• Updated the default clusterctl template to use external cloud-controller-manager (#4678, @nrb)
• Use seedling emoji for dependabot pull requests (#4652, @kranurag7)
• Wrap the underlying error when the AWSMachinePool reconciler fails to get the infra provider cluster or control plane object (#4685, @harveyxia)

The images for this release are: registry.k8s.io/cluster-api-aws/cluster-api-aws-controller:v2.4.0

Thanks to all our contributors.