v1.6.0

[gardener-extension-provider-aws]

Action Required

• [USER] The AWS extension does now support shoot clusters with Kubernetes version 1.18. You should consider the Kubernetes release notes before upgrading to 1.18. Please note that the AWS EBS CSI driver will be used for 1.18 shoots. It is compatible with the legacy volume provisioners, however, you might want to update your storage classes and volume handling accordingly. Please find more information about CSI in the official Kubernetes documentation. (#64, @rfranzke)

Improvements

• [OPERATOR] kube-controller-manageer no longer has access to blocked CIDRs. (#69, @mvladev)
• [OPERATOR] g/gardener-extensions does now require the g/terraformer@v1.0.0. (#57, @ialidzhikov)

[cloud-provider-aws]

Improvements

• [OPERATOR] k8s.io/legacy-cloud-providers is now updated to v1.18.0. (gardener-attic/cloud-provider-aws@87ca4e4)

[gardener-extensions]

Improvements

• [OPERATOR] Add Separate sidecar controller to save worker state under Worker.Status.State (gardener-attic/gardener-extensions#600, @vlvasilev)
  • Add Worker.Status.State restoration functionality
  • Add migration functionality
• [OPERATOR] The machine.sapcloud.io CRDs applied by the ./pkg/controller/worker package do now get a protection label for accidental deletion. See also: gardener/gardener#2066 (gardener-attic/gardener-extensions#625, @rfranzke)

[terraformer]

Action Required

• [OPERATOR] The Terraformer does now lookup the relevant data stored in ConfigMaps or Secrets live from the system instead of relying on mounted volumes. This is a breaking change as the volume mount approach does no longer work, please adapt your manifests according to the examples. The rationale behind it is to not rely on potentially stale kubelet cache while it mounts the volume which may, in rare cases, cause state loss. (gardener/terraformer#37, @rfranzke)

Improvements

• [OPERATOR] nsxt and random providers are now removed from the terraform bundle. (gardener/terraformer#38, @ialidzhikov)
• [OPERATOR] terraformer does no longer ignore the termination signals sent to PID 1. It does now send a termination signal to the terraform process itself and waits for its completion. This should prevent rare cases in which the terraformer was not storing the state of created infrastructure resources. (gardener/terraformer#36, @ialidzhikov)

Docker Images

gardener-extension-provider-aws: eu.gcr.io/gardener-project/gardener/extensions/provider-aws:v1.6.0
gardener-extension-validator-aws: eu.gcr.io/gardener-project/gardener/extensions/validator-aws:v1.6.0