Changes since v0.4.1
⚠️ Action Required
IAM policy
The IAM policy for the controllers requires a new permission, tag:GetResources. You must manually add this permission to your IAM policy. v0.4.3 adds this to clusterawsadm but it is missing from v0.4.1 and v0.4.2.
Leader election
The default leader election ID has been updated in this release. When upgrading from a previous v0.4.1 (or older) release you will need to:
- Scale down the old controllers:
kubectl -n capa-system scale deployment/controller-manager --replicas=0. - Deploy the updated manifests with the newer image.
- Scale up the new controllers:
kubectl -n capa-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.
✨ New Features
- AWSCluster and AWSMachine default concurrency (#1163)
- Apply elb-related tags to managed subnets (#1174)
🐛 Bug Fixes
- Delete cloud provider provisioned load balancers and security groups (#1168)
- Actually apply explicitly specified subnet tags (#1177)
- Correct IAM profile for control plane (#1182)
- Support running alongside other Cluster API pods in the same namespace with leader election enabled (#1178)
- Add well-known tolerations to released CAPA manager manifest (#1170)
The image for this release is us.gcr.io/k8s-artifacts-prod/cluster-api-aws/cluster-api-aws-controller:v0.4.2.
Thanks to all our contributors! 😊