v1.58.0

ACTION REQUIRED

• The field vmNetName is considered deprecated for vSphere and networks should be used instead (#1673)

Bugfixes

• Fix assignment of compute and sizing policies for VMware Cloud Director (#1685)
• VSphere: Fix a bug where datastore cluster value was not being propagated to the cloud-config (#1682)
• Fixed multiple instance creation issue for vultr by waiting for instance to get started (#1706)

Chore

• Machine Controller is now built using Go 1.21.0 (#1698)
• Upgrade govulr to v3 and handle go vultr API response codes (#1706)

Cleanup

• Remove support for centos with Hetzner (#1725)
• Remove support for Kubernetes 1.24 (#1675)
• Remove support for Kubernetes v1.25 (#1709)
• Support for container runtime docker has been removed (#1692)

New Feature

• Support for configuring multiple networks for vSphere (#1673)
• Anexia Provider: allow configuration of multiple disks (#1665)
• Kubevirt: support OCI VM image source (#1672)
• Machine-controller is now built using Go 1.21.3 (#1728)
• Machine-controller will inject provider-id for the machines that are created against cloud providers that don't have in-tree or external CCM support (#1723)
• Support for Kubernetes v1.28 (#1699)
• Add support for Vultr bare metal instances and Vultr VPC v1/v2 (#1706)

v1.57.3

What's Changed

• Propagate datastore cluster for vSphere to cloud-config #1695

Full Changelog: v1.57.2...v1.57.3

v1.57.2

What's Changed

• Fix version of k8s.io/client-go in go.mod to v0.26.4 in #1694

Full Changelog: v1.57.1...v1.57.2

v1.57.1

ACTION REQUIRED

• Support for configuring multiple networks for vSphere. The field vmNetName is considered deprecated for vSphere and networks should be used instead (#1673)

Bugfixes

• Fix assignment of compute and sizing policies for VMware Cloud Director (#1685)

New Feature

• Anexia Provider: allow configuration of multiple disks (#1665)

v1.56.4

Bugfixes

• Fix assignment of compute and sizing policies for VMware Cloud Director (#1685)

New Feature

• Anexia Provider: allow configuration of multiple disks (#1668)

v1.54.8

Bugfixes

• Fix assignment of compute and sizing policies for VMware Cloud Director (#1685)

New Feature

• Anexia Provider: allow configuration of multiple disks (#1668)

v1.57.0

ACTION REQUIRED

• The default values for webhook flags tls-cert-path and tls-key-path have been changed to /tmp/cert/tls.crt and /tmp/cert/tls.key respectively. This is a result of updating it to use certificates generated from cert-manager (#1644)
• Deprecation: User data plugins for machine-controller are deprecated and will soon be removed. Operating System Manager is the successor of user data plugins. It's responsible for creating and managing the required configurations for worker nodes in a Kubernetes cluster with better modularity and extensibility. Please refer to Operating System Manager for more details (#1642)

Bugfixes

• Allow external CCMs to handle failing node objects before MC (#1645)
• Anexia Provider: fix cleanup of failed machines (#1600)
• Anexia Provider: Fix nil pointer exception when providing no network configuration (#1646)
• Anexia Provider: fixed race condition in Cleanup method (#1571)
• Cancelled or timed out admission requests are not cached anymore (#1593)
• For generating cloud-config for Kubernetes 1.27 machines on AWS, make sure to set DisableStrictZoneCheck to false, as the flag is not supported in 1.27+ anymore (#1638)
• Machine deletion does not get stuck if assigned floating ip is absent (e.g. deleted externally) (#1616)
• Reduce default log verbosity from 6 to 4 (#1597)

New Feature

• Add machine_controller_provisioning_time_seconds and machine_controller_deprovisioning_time_seconds metrics to the machine controller (#1572)
• Add support for Kubernetes v1.27 (#1623)
• Add support for OpenNebula as a cloud provider (#1450)
• Add VMs to an Anti-Affinity rule (vSphere only). This requires additional permissions: (#1647)
  • AutoDeploy
    • Rule
      • Create
      • Delete
      • Edit
  • Host
    • Inventory
      • Modify cluster
• Anexia Provider: Enable up to 4 custom dns entries (#1637)
• AWS: Machines using a VPC without enableDnsHostname enabled will now fail validation, as those machines would not be able to register themselves successfully as nodes in the cluster (#1590)
• Logging was replaced with Zap; the klog-related CLI flags like -logtostderr, -v etc. have been removed; use -log-debug=true/false and -log-format=text/json instead (#1606)
• Support for finding nodes by providerID for VMware Cloud Director, Nutanix, and linode (#1568)
• Support for Vultr cloud provider (#1531)
• Support Public IPs assignments for Hetzner cloud (#1579)
• Create the file /etc/iscsi/initiatorname.iscsi, if not present, start and enable the iscsid.service for any new VM, if you use flatcar nodes on Nutanix (#1596)
• VMware Cloud Director now supports authentication using API Token (#1612)

Chore

• Default timeout for OpenStack client has been increased to 300 seconds (#1570)
• Update to Go 1.20.5 (#1659)

Documentation

• Clarify provider status for Linode, Vultr and OpenNebula as community providers (#1631)

v1.56.3

Bugfixes

• Allow external CCMs to handle failing node objects before MC (#1652)
• Anexia Provider: Fix nil pointer exception when providing no network configuration (#1650)

New Feature

• Anexia Provider: Enable up to 4 custom dns entries (#1640)

Misc

• Machine-controller is now built with Go 1.19.9 (#1634)

v1.54.7

Bugfixes

• Allow external CCMs to handle failing node objects before MC (#1654)
• Anexia Provider: Fix nil pointer exception when providing no network configuration (#1649)

New Feature

• Anexia Provider: Enable up to 4 custom dns entries (#1641)

Misc

• Machine-controller is now built with Go 1.18.10 (#1635)

v1.56.2

What's Changed

• [release/v1.56] Bump Golang to 1.19.8 by @pkprzekwas in #1615

Full Changelog: v1.56.1...v1.56.2