v1.20.2

• Fix openstack instanceReadyPeriod/Timeout (#848)

v1.20.1

• Refactor the KubeVirt DNS settings (#846)
• Mark Anexia cloud provider as an alpha provider (#845)

v1.20.0

• Add support for Anexia Engine provider (#842)
• Add support for Scaleway provider (#835)
• Support cloud-init on Flatcar (#840)
• instanceReadyCheckPeriod and instanceReadyCheckTimeout are now configurable for OpenStack Machines to allow alternative timeouts and retries. 120s is the default (#843)

v1.19.0

Note: The Poseidon Kubelet repository that's used for Flatcar instances doesn't publish 1.18 images any longer. This machine-controller release reverts to the Hyperkube Kubelet image for 1.18 clusters, as it's still available for releases up to 1.19.

• Use Hyperkube Kubelet instead of Poseidon Kubelet for 1.18 clusters (#837)
• Fix the polkit rule filename to allow core user to reboot Flatcar machines for auto-updates (#830)
• Always ensure to have machine-delete-finalizer (#811)

v1.18.0

Note: CoreOS images have got removed on AWS, GCP, and DigitalOcean. If you're still running CoreOS on those providers, you'll need to build your own CoreOS image.

• Add support for setting arbitrary feature gates on the kubelet (#821)
• Fix the Docker installation on RHEL8 (#826)
• Fix hanging reboot on auto-update on Flatcar Linux (#824)

v1.14.4

• Fix the Docker installation on RHEL8 (#826)

v1.17.2

• new hetzner field "image" (#817)
• Update CNI plugins to v0.8.7 (#818)

v1.17.1

• Ignore Node when deleting by NodeRef if it doesn't exist anymore (#816)

v1.17.0

• Update Docker to 19.03.12 for Ubuntu, CentOS, and RHEL clusters running Kubernetes 1.17 and newer (#815)
• Delete the Node object by NodeRef if it's present (#813)
• Compare node and instances names when applying the NodeOwner label for Hetzner machines/instances (#813)
• Support Flatcar for Packet cloud provider (#802)

v1.16.1

• Fix systemd setup.service startup after reboot (#810)