Design of Machine Controller Manager
The design of the Machine Controller Manager is influenced by the Kube Controller Manager, where-in multiple sub-controllers are used to manage the Kubernetes clients.
It's designed to run in the master plane of a Kubernetes cluster. It follows the best principles and practices of writing controllers, including, but not limited to:
Reusing code from kube-controller-manager
leader election to allow HA deployments of the controller
workqueues and multiple thread-workers
SharedInformers that limit to minimum network calls, de-serialization and provide helpful create/update/delete events for resources
rate-limiting to allow back-off in case of network outages and general instability of other cluster components
sending events to respected resources for easy debugging and overview
Prometheus metrics, health and (optional) profiling endpoints
Objects of Machine Controller Manager
Machine Controller Manager makes use of 4 CRD objects and 1 Kubernetes secret object to manage machines. They are as follows,
Machine-class: Represents a template that contains cloud provider specific details used to create machines.
Machine: Represents a VM which is backed by the cloud provider.
Machine-set: Represents a group of machines managed by the Machine Controller Manager.
Machine-deployment: Represents a group of machine-sets managed by the Machine Controller Manager to allow updating machines.
Secret: Represents a kubernetes secret that stores cloudconfig (initialization scripts used to create VMs) and cloud specific credentials
Components of Machine Controller Manager
Machine Controller Manager is made up of 3 sub-controllers as of now. They are -
Machine Controller: Used to create/update/delete machines. It is the only controller which actually talks to the cloud providers.
Machine Set Controller: Used to manage machine-sets. This controller makes sure that desired number of machines are always up and running healthy.
Machine Deployment Controller: Used to update machines from one version to another by manipulating the machine-set objects.
Machine Safety Controller: A safety net controller that terminates orphan VMs and freezes machineSet/machineDeployment objects which are overshooting or timing out while trying to join nodes to the cluster.
All these controllers work in an co-operative manner. They form a parent-child relationship with Machine Deployment Controller being the grandparent, Machine Set Controller being the parent, and Machine Controller being the child.
The following is a short list of future plans,
Integrate the cluster-autoscaler to act upon machine-deployment objects, used to manage the required number of machines based on the load of the cluster.
Support other cloud providers like OpenStack.
Integrate a garbage collector to terminate any orphan VMs.
Build a comprehensive testing framework.
Fix bugs that exist in the current implementation.
This link contains the working doc for the todos which are planned in the near future.