v2.7.1

[richardcase]

Release notes for Cluster API Provider AWS (CAPA) v2.7.1

Documentation

Changelog since v2.6.1

NOTE: there is no v2.7.0 release available due to an issue encountered during the release process.

Urgent Upgrade Notes

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

• A new community owned AWS account is now being used for publishing AMIs. The old account owner by a company is no longer accessible and the images are being automatically deleted by AWS. Support for ubuntu 18.04 & 20.04 has been dropped and ubuntu 24.04 is now the default OS if no OS is supplied. (✨refactor: update to use new community ami account #5133, @richardcase)
• View the projects policy on publishing AMIs here.
• There is known issue creating clusters with Kubernetes v1.30 or greater when the load balancer is an ELB (which is the default). Its recommended for new clusters that you use an NLB instead, See ELB health check fails with Kubernetes >=v1.30.x #5139 for further details.

Changes by Kind

Support

• Add issue templates for reviewer and maintainer changes (🌱 feat: add issue templates for reviewers/maintainer changes #5079, @richardcase)
• Bump CAPI to v1.8.4 (🌱 Bump CAPI to v1.8.4 #5061, @damdo)
• Switch PodDisruptionBudget in templates to policy/v1 (🌱 Switch PodDisruptionBudget in templates to policy/v1 #5102, @AndiDog)
• Templates: align cluster-template.yaml and cluster-template-machinepool.yaml, use external cloud controller and CSI (🌱 Align cluster-template.yaml and cluster-template-machinepool.yaml, use external cloud controller and CSI #5136, @AndiDog)
• Update metadata file for v2.7.0 release. (🚀chore: update metadata for release 2.7.0 #5187, @richardcase)
• Update the metadata file for v2.6.0 (🌱chore: update metadata for v2.6.0 #5071, @richardcase)

Feature

• Add NonRootVolumes to AWSMachinePool launch template. (✨ Add non root volumes to AWSMachinePool #4954, @mnitchev)
• Add ROSA Control plane registry configuration. (✨ Add cluster registry config #5132, @serngawy)
• Change prevenet deletion label to be specific to CAPA. (🌱 chore: change name of prevent deletion label #5072, @richardcase)
• Feat: add support for additionalControlPlaneIngressRules in AWSManagedControlPlane (✨ Add support for additionalControlPlaneIngressRule on AWSManaged… #4783, @fad3t)
• New AMI building workflow that allows a packer vars file to be passed in. (🌱 feat: build amis with vars file #5152, @richardcase)
• New AWSCluster.spec.network.NodePortIngressRuleCidrBlocks parameter which allows to specify IPv4 CIDR blocks for Security group rule "Node Port Services" instead of default 0.0.0.0/. (✨ CIDR blocks for default Security Group rule "Node Port Services" #5147, @krasoffski)
• New workflow to publish CAPA AMIs (✨ chore: workflow to build and publish AMIs #5138, @richardcase)
• Support adding custom secondary VPC CIDR blocks in AWSCluster (✨ Support adding custom secondary VPC CIDR blocks in AWSCluster #4898, @AndiDog)
• Support setting maxHealthyPercentage to configure ASG instance refresh (✨ Support setting maxHealthyPercentage to configure ASG instance refresh #5140, @fiunchinho)

Bug or Regression

• A missing permission needed for using nlbs (🐛 fix: add missing permissions for nlb #5154, @richardcase)
• Check for the LB "active" status instead of trying to resolve the DNS name to validate the LB is ready. (🐛: elbv2: wait for LB active state instead of resolving DNS name #5093, @r4f4)
• Fix duplicated/leaked EIP when using BYO IPv4 on Machines. (🐛 ec2/byoip: fix EIP leak when creating machine #5039, @mtulio)
• Fix kustomize issue with e2e tests. (🐛 fix: kustomize problem in e2e #5167, @richardcase)
• Fix: do not attach security groups for Network Load Balancers in secret regions. (🐛: elbv2: skip adding security groups to NLB in secret regions #5030, @r4f4)
• Mark S3BucketCreated condition ready after successful reconciliation (🐛 Mark S3BucketCreated condition ready after successful reconciliation #5089, @AndiDog)
• Updates EKS e2e tests after recent bumping of EKS versions. (🌱fix: eks tests broken due to incorrect addon versions #5185, @richardcase)

Uncategorized

• Add versionGate Acknowledgement for upgrading ROSA-HCP y-stream versions (🐛 ROSA: Fix version gate Ack for rosa hcp y-stream version upgrade #5166, @serngawy)
• Fix: Fargate: Standardize cluster name in role names to avoid errors on mismatch between Cluster CR and EKS cluster name (🐛 Standardize cluster name in Fargate role names to avoid errors on mismatch between Cluster CR and EKS cluster name #5111, @alam0rt)
• Update Flatcar template to use external CCM. ( 🐛 Use external CCM in Flatcar template #5099, @ader1990)

The images for this release are:

registry.k8s.io/cluster-api-aws/cluster-api-aws-controller:v2.7.1

Thanks to all our contributors.

---

This discussion was created from the release v2.7.1.

[faermanj]

It seems that cluster-template-development.yaml is missing on this release

Error: failed to read "cluster-template-development.yaml" from provider's repository "infrastructure-aws": failed to download files from GitHub release v2.7.1: failed to get file "cluster-template-development.yaml" from "v2.7.1" release

[damdo]

I went back to the v2.1 release and looked for cluster-template-development.yaml.
I don't think we ever had such file.
What component is throwing that error, is it clusterctl?
Is it configured with some special development flags or environment or flavour?

[faermanj]

Found this while trying to get through the quickstart.
I'll share a better review.

[richardcase]

If you click on "Other Providers" instead and run that command you should be good.

[richardcase]

So this:

clusterctl generate cluster capi-quickstart \
  --kubernetes-version v1.31.0 \
  --control-plane-machine-count=3 \
  --worker-machine-count=3 \
  > capi-quickstart.yaml

NOTE: there is no --flavor flag