OpenShift Container Platform

Deploying user-defined firmware to cluster hosts with GitOps ZTP

You can deploy user-defined firmware settings to cluster hosts by creating custom node templates that include HostFirmwareSettings CRs, and referencing them in the ClusterInstance CR. You can configure hardware profiles to apply to hosts in the following scenarios:

All hosts in the cluster
Individual hosts in the cluster

You can configure host hardware profiles to be applied in a hierarchy. Node-level profiles override cluster-wide settings.

Prerequisites

You have installed the OpenShift CLI (oc).
You have installed Red Hat Advanced Cluster Management (RHACM) version 2.12 or later and logged in to the hub cluster as a user with cluster-admin privileges.
You have installed the SiteConfig Operator in the hub cluster.
You created a Git repository where you manage your custom site configuration data. The repository must be accessible from the hub cluster and be defined as a source repository for the Argo CD application.

Procedure

Create the HostFirmwareSettings CR that contains the firmware settings you want to apply. For example, create the following YAML file:

host-firmware-settings.yaml

apiVersion: metal3.io/v1alpha1
kind: HostFirmwareSettings
metadata:
  name: "site1-sno-du"
  namespace: "site1-sno-du"
spec:
  settings:
    BootMode: "Uefi"
    LogicalProc: "Enabled"
    ProcVirtualization: "Enabled"

Save the HostFirmwareSettings CR file relative to the kustomization.yaml file that you use to provision the cluster. For example:

site-configs/
  └── site1-sno-du/
        ├── clusterinstance-site1-sno-du.yaml
        ├── kustomization.yaml
        └── host-firmware-settings.yaml

Create a ConfigMap to store the HostFirmwareSettings CR. You can use a kustomization.yaml file with configMapGenerator to create the ConfigMap. For example:

apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
  - clusterinstance-site1-sno-du.yaml
configMapGenerator:
  - name: host-firmware-settings-cm
    namespace: site1-sno-du
    files:
      - host-firmware-settings.yaml
generatorOptions:
  disableNameSuffixHash: true

namespace must match the ClusterInstance namespace.
host-firmware-settings.yaml is the name of the HostFirmwareSettings CR.

To apply a hardware profile to all hosts in the cluster, reference the ConfigMap in the spec.templateRefs field of your ClusterInstance CR. For example:

apiVersion: siteconfig.open-cluster-management.io/v1alpha1
kind: ClusterInstance
metadata:
  name: "site1-sno-du"
  namespace: "site1-sno-du"
spec:
  clusterName: "site1-sno-du"
  # ...
  templateRefs:
    - name: host-firmware-settings-cm
      namespace: site1-sno-du
  nodes:
    - hostName: "node1.example.com"
      # ...

templateRefs applies the firmware profile to all hosts in the cluster.

Optional: To apply a hardware profile to a specific host in the cluster, reference the ConfigMap in the spec.nodes[].templateRefs field. For example:

apiVersion: siteconfig.open-cluster-management.io/v1alpha1
kind: ClusterInstance
metadata:
  name: "site1-sno-du"
  namespace: "site1-sno-du"
spec:
  clusterName: "site1-sno-du"
  # ...
  nodes:
    - hostName: "node1.example.com"
      # ...
      templateRefs:
        - name: host-firmware-node1-cm
          namespace: site1-sno-du
    - hostName: "node2.example.com"
      # ...

nodes[].templateRefs applies the firmware profile only to the node1.example.com host.

Node-level templateRefs settings override cluster-level templateRefs settings.

Commit the ClusterInstance CR, ConfigMap, and associated kustomization.yaml changes in your Git repository and push the changes.

The Argo CD pipeline detects the changes and begins the managed cluster deployment.

Cluster deployment proceeds even if an invalid firmware setting is detected. To apply a correction using GitOps ZTP, re-deploy the cluster with the corrected hardware profile.

Verification

Check that the firmware settings have been applied in the managed cluster host. For example, run the following command:

$ oc get hfs -n <managed_cluster_namespace> <managed_cluster_name> -o jsonpath='{.status.conditions[?(@.type=="Valid")].status}'

<managed_cluster_namespace> is the namespace of the managed cluster and <managed_cluster_name> is the name of the managed cluster.

The following example output shows valid firmware settings:
```
True
```