With Red Hat OpenShift Container Platform 3, you are usually working with three roles for nodes: master, compute, and infrastructure (infra). Red Hat OpenShift Container Platform 4 instead, by default, provides master and worker roles. There is no infra node in OpenShift 4. When migrating from OpenShift 3 to OpenShift 4, users frequently ask how to convert OpenShift 3 infra nodes to OpenShift 4 infra nodes. In this post we'll walk you through that.

Red Hat ausgezeichnet als Leader im 2023 Gartner® Magic Quadrant™

Red Hat wurde im Gartner 2023 Magic Quadrant für Container Management hinsichtlich der Ausführungsfähigkeit und der umfassenden Vision bestplatziert.

The most common mistake that people make is just applying the infra label to the worker nodes:

oc label node $NODE_NAME node-role.kubernetes.io/infra=

After applying the label, there is no output to suggest that this is a problem. However, when you upgrade the OpenShift cluster, you will notice that the worker nodes that you applied the label to do not get upgraded. This is the expected behavior.

In this post we will discuss what Machine Config Pool is and how you should use it for OpenShift 4 infra nodes.

Bottom line

In order to understand the situation, you should understand the concept of Operators in OpenShift Container Platform 4.

As you may know, OpenShift Container Platform 4 is an operator-focused platform. On Red Hat CoreOS, the Machine Config Operator handles the operating system, managing OS updates and configuration changes. Understanding the Machine Config Operator is central to managing master and worker nodes properly according to the new management design of OpenShift Container Platform 4.

What is a Machine Config Operator? 

The Machine Config Operator (MCO) is a cluster-level operator like any other operator, but it is a special one from an OpenShift Container Platform infrastructure perspective. It manages the operating system and keeps the cluster up to date and configured. Through MCO, platform administrators can configure and update systemd, cri-o/kubelet, kernel, NetworkManager, etc. on the nodes. To do so, the MCO creates a statically rendered MachineConfig file which includes the MachineConfigs for each node. It then applies that configuration to each node. Let’s examine some of the details.

Sub Components of Machine Config Operator

The Machine Config Operator is a complex component. There are several sub-components and each sub-component performs a different task. This blog does not explain all the sub-components in detail, but you can examine each of them by clicking the links below.

Containers illustrationErreichen Sie mehr mit dem Universal Base Image (UBI) von Red Hat

What does Machine Config Controller do?

Of the above sub-components, the Machine Config Controller is relevant for the problem that we are discussing. The Machine Config Controller starts to play a role after OpenShift Container Platform 4 is installed.

Machine Config Controller Goals:

  • Coordinate upgrade of machines to the desired configuration defined by a MachineConfig Object.

  • Provide options to control upgrades for sets of machines individually.

In other words, this controller generates Machine Configs for pre-defined roles (master and worker) and monitors whether an existing Machine Config CR (custom resource) is modified or new Machine Config CRs are created. When the controller detects any of those events, it will generate a new rendered Machine Config object that contains all of the Machine Configs based on MachineConfigSelector from each MachineConfigPool.

The following are each controller’s role description:

Template Controller:

  • Generate the MachineConfigs for predefined roles of machines (master, worker).

  • Watch controllerconfig to generate OpenShift-owned MachineConfigs.

Render Controller:

  • Watch MachineConfigPool object to find all the MachineConfig objects. 

  • Update CurrentMachineConfig with the rendered MachineConfig.

  • Detect changes on all the MachineConfigs and syncs all the MachineConfigPool objects with a new CurrentMachineConfig.

Update Controller:

  • Watch if MachineConfigPool .Status.CurrentMachineConfig is updated.

  • Upgrade machines to the desired MachineConfig by coordinating with a daemon running on each machine.

Machine Config

We have mentioned the Machine Config object, but what is it? It is the source of machine configuration and is used for installation and first-boot, as well as upgrades. The Machine Config should be static, i.e. it should not contain links or remote locations to generate configuration dynamically. In short, it defines OS level configuration if the machine uses RHCOS. Using the ignition config format, you can manage machines such as storage, systemd, kernel and so on.

Machine Config Pool 

The Machine Config Pool is the main object to resolve the situation. The Machine Config Pool operates similarly to the Rolebinding object, which associates roles with users. The Machine Config Pool associates nodes with Machine Configs.

Let’s examine the following diagram to understand the Machine Config Pool. 

The Machine Config Pool maps between nodes and Machine Configs. The Machine Config Pool has two selectors, and each selector matches machine configs with nodes.

 

MachineConfigPool API diagram

Figure 1. The relationship among Machine Config Pool/Machine Configs and Worker Nodes 

The Render controller in the Machine Config Controller monitors the Machine Config Pool and generates static machine config objects named rendered-master-XXXX and rendered-worker-xxx. These objects can include multiple machine configs. The Render controller then checks whether the nodes in the pool have applied the latest rendered-xxxx machine config. If the machine config pool changes, then the render controller creates a new rendered-xxx and applies it.

 

Figure 2. How does Machine Config Pool select Machine Configs and Worker

Figure 2. How Machine Config Pool selects Machine Configs and Worker

Infra worker node group

Now it is time to solve the infra role worker nodes issue. As I mentioned, there are only 2 roles in OpenShift 4 by default: master and worker. Therefore, the OpenShift 3 infra role needs to inherit the OpenShift 4 worker role in order for the upgrade to take place.

Step 1. Create a infra MachineConfigPool

apiVersion: machineconfiguration.openshift.io/v1
kind: MachineConfigPool
metadata:
  name: infra
spec:
  machineConfigSelector:
    matchExpressions:
      - {key: machineconfiguration.openshift.io/role, operator: In, values: [worker,infra]}
  maxUnavailable: null
  nodeSelector:
    matchLabels:
      node-role.kubernetes.io/infra: ""
  paused: false

From MachineConfigSelector, you can see it tries to find worker and infra related Machine Configs. There are no infra related MachineConfigs by default but if you want to configure something only for infra worker nodes, you can create it according to Step 2.

Step 2. Create a infra MachineConfig (optional)

apiVersion: machineconfiguration.openshift.io/v1
kind: MachineConfig
metadata:
  labels:
    machineconfiguration.openshift.io/role: infra
  name: 50-infra
spec:
  config:
    ignition:
      version: 2.2.0
    storage:
      files:
      - contents:
          source: data:,test
        filesystem: root
        mode: 0644
        path: /etc/testinfra

Step 3. Update worker node role of the infra nodes.

oc label node <node> node-role.kubernetes.io/worker-

oc label node <node> node-role.kubernetes.io/infra=

Conclusion

Red Hat OpenShift Container Platform 4 includes many operators and it is managed by them from installation, upgrade and so on. Unlike OpenShift Container Platform 3, you should follow the new ways of OpenShift 4 to manage worker nodes by Machine Config. Machine Config Operator is a main component to control and handle OpenShift worker nodes so you should understand this concept and use it properly for OpenShift Container Platform 4.


About the author

Jooho Lee is a senior OpenShift Technical Account Manager (TAM) in Toronto supporting middleware products(EAP/ DataGrid/ Web Server) and cloud products (Docker/ Kubernetes/ OpenShift/ Ansible). He is an active member of JBoss User Group Korea and Openshift / Ansible Group. 

Read full bio