v0.29.0

[machine-controller-manager]

Improvements

• [USER] Azure: Add support for custom image plans for VMs (#463, @MSSedusch)

v0.28.0

[machine-controller-manager]

Improvements

• [USER] Provide a way to specify "auth-extra-groups" field in created bootstrap tokens. (#442, @zuzzas)
• [OPERATOR] The drain is completely skipped if the node is NotReady for the timeout period. Default timeout is 5 mins. (#450 , @zuzzas)
• [USER] Support multiple network interfaces in the AWS driver (#441, @zuzzas)
• [OPERATOR] The mcm stopped adopting Nodes with empty ProviderID field (#451, @zuzzas)
• [OPERATOR] Bugfix: Continue with VM deletion when data disk doesn't exist for AWS. (#447, @prashanth26)
• [OPERATOR] Add GCP local ssd disk support (#414, @tennix)
• [DEVELOPER] Added support for multiple Data Disks in Azure and AliCloud. (#397, @guydaichs)

v0.27.0

[machine-controller-manager]

Most notable changes

• [OPERATOR] Prepend mcm to all work queue metrics (#420, @prashanth26)
• [OPERATOR] Subsystems and Namespaces to MCM metrics (#420, @prashanth26)
• [OPERATOR] Renamed mcm_machine_deployment_items_total & mcm_machine_set_items_total metrics (#420, @prashanth26)
• [OPERATOR] Bugfix: Set deleteOnTermination to true by default for volumes. Disks that are created with the instance are deleted with instance termination. (#420, @prashanth26)

Improvements

• [USER] Azure: Support for accelerated networking enabled VMs (#438, @dkistner)
• [OPERATOR] Azure: VMs now can be created with image ID. (#436, @vpnachev)
• [OPERATOR] Improves the machine-creation and machine-deletion flow by listing and verifying the state of the machines before actual creation and deletion. (#432, @hardikdr)
• [OPERATOR] MCM now dynamically maps the node-objects with machines if Status.Node is not set. (#432, @hardikdr)
• [OPERATOR] Sort active machines by creation timestamp (#429, @prashanth26)
• [OPERATOR] Openstack: When a Server creation ERRORs, provide formatted Fault information (#427, @zuzzas)
• [OPERATOR] Set deleteOnTermination flag to true while deleting the VMs in AWS. (#420, @prashanth26)
• [OPERATOR] Skip eviction during drain if ForceDeletePods is enabled. (#416, @amshuman-kr)
• [OPERATOR] Openstack: Print API request/response debug when verbosity is set to 6 (#412, @kayrus)
• [OPERATOR] Azure: VM deletion checks any orphan dependant resources and tries to delete them if exists. (#406, @dkistner)
• [DEVELOPER] Azure: Update of the Azure SDK to use more recent version (#406, @dkistner)

v0.26.3

[machine-controller-manager]

Improvements

• [OPERATOR] Set deleteOnTermination flag to true while deleting the VMs in AWS. (#409, @prashanth26)

v0.26.2

[machine-controller-manager]

Most notable changes

• [OPERATOR] Bugfix: Set deleteOnTermination to true by default for volumes. Disks that are created with the instance are deleted with instance termination. (#408, @hardikdr)

v0.26.1

[machine-controller-manager]

Most notable changes

• [OPERATOR] Prepend mcm to all work queue metrics (#405, @prashanth26)
• [OPERATOR] Subsystems and Namespaces to MCM metrics (#405, @prashanth26)
• [OPERATOR] Renamed mcm_machine_deployment_items_total & mcm_machine_set_items_total metrics (#405, @prashanth26)

v0.26.0

[machine-controller-manager]

Most notable changes

• [USER] Updated dependecies to refer k8s 1.16.0 (#401, @prashanth26)
• [OPERATOR] Promethus worker metrics have aligned with k8s version 1.16 (#401, @prashanth26)
• [OPERATOR] Healthz endpoint is immediately initialized with a healty (200) response (#375, @zuzzas)

Improvements

• [USER] Azure machines can now be deployed with an attached user assigned managed identity. (#395, @dkistner)
• [USER] Allows specifying multiple networks for a VM via the new OpenstackMachineClassSpec "networks" field (#376, @zuzzas)
• [OPERATOR] Stops reconciling machine objects that no longer exists (#403, @ggaurav10)
• [OPERATOR] Bugfix: Fixes race between machine & machineSet creation/deletion operations (#391, @prashanth26)
• [OPERATOR] Added Cinder based root disk support with customisable disk size (#388, @afritzler)
• [OPERATOR] Allow status to be defined when replicas is not defined or equal to 0. (#387, @rewiko)
  • category: improvement
  • target_group: user
• [OPERATOR] Added support for OpenStack machine image id. (#374, @afritzler)
• [OPERATOR] Avoided force-deleting the machine if previous drain has failed. (#373, @hardikdr)
• [OPERATOR] Drain logic now attempts to evict the pod till drain-timeout has occurred. The interval between consecutive attempts to evict the pod has been increased to 20s. (#370, @hardikdr)
• [OPERATOR] github.com/Azure/azure-sdk-for-go is updated to v32.6.0. (#369, @ialidzhikov)
• [OPERATOR] bootstrap token generation is handled at MCM (#351, @vlvasilev)
• [DEVELOPER] Fixed license header (#385, @afritzler)

v0.25.0

[machine-controller-manager]

Improvements

• [OPERATOR] AliCloud: When nodes are rolling updated, pods with Persistent volume attached can now be evicted correctly/serially before the node is drained. (#366, @jia-jerry)
• [OPERATOR] Bugfix: Remove dependency on MachineClass while deleting MachineDeployment/MachineSets (#365, @prashanth26)
• [OPERATOR] Bugfix: Allow deletion of machine without status.currentStatus.phase (#364, @prashanth26)
• [OPERATOR] The release tags from now are prefixed with v. (#362, @ialidzhikov)
• [OPERATOR] OpenStack: Ensure VMs are deleted after unsuccessful create operation (#356, @kayrus)
• [DEVELOPER] The golang version has been updated to 1.13. Please upgrade your local go installation to 1.13. (#360, @ialidzhikov)

0.24.0

[machine-controller-manager]

Action Required

• [USER] Since version 0.22.0 the fields Spec.Properties.StorageProfile.ImageReference.[Publisher|Offer|Sku|Version] in the AzureMachineClass have been deprecated and now they are completely removed. Please switch to Spec.Properties.StorageProfile.ImageReference.URN` before upgrading to this version or higher. (#353, @vpnachev)

Improvements

• [OPERATOR] Node-conditions are configurable via flag. Users can provide the list of node-conditions based on which machines are replaced if set to true for health-timeout period. (#348, @hardikdr)
• [OPERATOR] Deletes any Azure infra resources before returning error on machine create (#347, @prashanth26)

0.23.0

[machine-controller-manager]

Most notable changes

• [OPERATOR] ⚠️ GCP VM's Public IP attachment is now optional. Default behaviour is that public IP attachment is enabled. To disable this, refer to this example here (#340, @prashanth26)

Improvements

• [USER] Create and manage Azure virtual machines assigned to a virtual network (vNet), which is located in a different resource group than the machines. (#344, @dkistner)
• [OPERATOR] Updated Integration Tests kubectl to v1.16.0 (#342, @prashanth26)
• [OPERATOR] Fixes leaks in Go Routines leading to CPU wastage (#341, @prashanth26)
• [DEVELOPER] Added multi stage docker build (#343, @afritzler)