This is a cache of https://docs.openshift.com/container-platform/4.17/edge_computing/ztp-using-hub-cluster-templates.html. It is a snapshot of the page at 2024-11-23T07:38:56.324+0000.
Using hub templates in PolicyGenerator or PolicyGenTemplate CRs | Edge computing | OpenShift Container Platform 4.17
    OpenShift docs are moving and will soon only be available at docs.redhat.com, the home of all Red Hat product documentation. Explore the new docs experience today.
  1. Documentation
    2. history bug_report picture_as_pdf
×

Using hub templates in PolicyGenerator or PolicyGenTemplate CRs

Topology Aware Lifecycle Manager supports Red Hat Advanced Cluster Management (RHACM) hub cluster template functions in configuration policies used with GitOps Zero Touch Provisioning (ZTP).

Hub-side cluster templates allow you to define configuration policies that can be dynamically customized to the target clusters. This reduces the need to create separate policies for many clusters with similiar configurations but with different values.

Policy templates are restricted to the same namespace as the namespace where the policy is defined. This means you must create the objects referenced in the hub template in the same namespace where the policy is created.

Using PolicyGenTemplate CRs to manage and deploy polices to managed clusters will be deprecated in an upcoming OpenShift Container Platform release. Equivalent and improved functionality is available using Red Hat Advanced Cluster Management (RHACM) and PolicyGenerator CRs.

For more information about PolicyGenerator resources, see the RHACM Policy Generator documentation.
Additional resources

Specifying group and site configurations in group PolicyGenerator or PolicyGentemplate CRs

You can manage the configuration of fleets of clusters with configmap CRs by using hub templates to populate the group and site values in the generated policies that get applied to the managed clusters. Using hub templates in site PolicyGenerator or PolicyGentemplate CRs means that you do not need to create a policy CR for each site.

You can group the clusters in a fleet in various categories, depending on the use case, for example hardware type or region. Each cluster should have a label corresponding to the group or groups that the cluster is in. If you manage the configuration values for each group in different configmap CRs, then you require only one group policy CR to apply the changes to all the clusters in the group by using hub templates.

The following example shows you how to use three configmap CRs and one PolicyGenerator CR to apply both site and group configuration to clusters grouped by hardware type and region.

There is a 1 MiB size limit (Kubernetes documentation) for configmap CRs. The effective size for the configmap CRs is further limited by the last-applied-configuration annotation. To avoid the last-applied-configuration limitation, add the following annotation to the template configmap:

argocd.argoproj.io/sync-options: Replace=true
Prerequisites

  • You have installed the OpenShift CLI (oc).

  • You have logged in to the hub cluster as a user with cluster-admin privileges.

  • You have 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 GitOps ZTP ArgoCD application.

Procedure

  1. Create three configmap CRs that contain the group and site configuration:

    1. Create a configmap CR named group-hardware-types-configmap to hold the hardware-specific configuration. For example:

      apiVersion: v1
kind: configmap
metadata:
  name: group-hardware-types-configmap
  namespace: ztp-group
  annotations:
    argocd.argoproj.io/sync-options: Replace=true (1)
data:
  # SriovNetworkNodePolicy.yaml
  hardware-type-1-sriov-node-policy-pfNames-1: "[\"ens5f0\"]"
  hardware-type-1-sriov-node-policy-pfNames-2: "[\"ens7f0\"]"
  # PerformanceProfile.yaml
  hardware-type-1-cpu-isolated: "2-31,34-63"
  hardware-type-1-cpu-reserved: "0-1,32-33"
  hardware-type-1-hugepages-default: "1G"
  hardware-type-1-hugepages-size: "1G"
  hardware-type-1-hugepages-count: "32"
      1 The argocd.argoproj.io/sync-options annotation is required only if the configmap is larger than 1 MiB in size.

    2. Create a configmap CR named group-zones-configmap to hold the regional configuration. For example:

      apiVersion: v1
kind: configmap
metadata:
  name: group-zones-configmap
  namespace: ztp-group
data:
  # ClusterLogForwarder.yaml
  zone-1-cluster-log-fwd-outputs: "[{\"type\":\"kafka\", \"name\":\"kafka-open\", \"url\":\"tcp://10.46.55.190:9092/test\"}]"
  zone-1-cluster-log-fwd-pipelines: "[{\"inputRefs\":[\"audit\", \"infrastructure\"], \"labels\": {\"label1\": \"test1\", \"label2\": \"test2\", \"label3\": \"test3\", \"label4\": \"test4\"}, \"name\": \"all-to-default\", \"outputRefs\": [\"kafka-open\"]}]"

    3. Create a configmap CR named site-data-configmap to hold the site-specific configuration. For example:

      apiVersion: v1
kind: configmap
metadata:
  name: site-data-configmap
  namespace: ztp-group
data:
  # SriovNetwork.yaml
  du-sno-1-zone-1-sriov-network-vlan-1: "140"
  du-sno-1-zone-1-sriov-network-vlan-2: "150"

    Each configmap CR must be in the same namespace as the policy to be generated from the group PolicyGenerator CR.

  2. Commit the configmap CRs in Git, and then push to the Git repository being monitored by the Argo CD application.

  3. Apply the hardware type and region labels to the clusters. The following command applies to a single cluster named du-sno-1-zone-1 and the labels chosen are "hardware-type": "hardware-type-1" and "group-du-sno-zone": "zone-1":

    $ oc patch managedclusters.cluster.open-cluster-management.io/du-sno-1-zone-1 --type merge -p '{"metadata":{"labels":{"hardware-type": "hardware-type-1", "group-du-sno-zone": "zone-1"}}}'

  4. Depending on your requirements, Create a group PolicyGenerator or PolicyGentemplate CR that uses hub templates to obtain the required data from the configmap objects:

    1. Create a group PolicyGenerator CR. This example PolicyGenerator CR configures logging, VLAN IDs, NICs and Performance Profile for the clusters that match the labels listed the under policyDefaults.placement field:

      ---
apiVersion: policy.open-cluster-management.io/v1
kind: PolicyGenerator
metadata:
    name: group-du-sno-pgt
placementBindingDefaults:
    name: group-du-sno-pgt-placement-binding
policyDefaults:
    placement:
        labelSelector:
            matchExpressions:
                - key: group-du-sno-zone
                  operator: In
                  values:
                    - zone-1
                - key: hardware-type
                  operator: In
                  values:
                    - hardware-type-1
    remediationAction: inform
    severity: low
    namespaceSelector:
        exclude:
            - kube-*
        include:
            - '*'
    evaluationInterval:
        compliant: 10m
        noncompliant: 10s
policies:
    - name: group-du-sno-pgt-group-du-sno-cfg-policy
      policyAnnotations:
        ran.openshift.io/ztp-deploy-wave: "10"
      manifests:
        - path: source-crs/ClusterLogForwarder.yaml
          patches:
            - spec:
                outputs: '{{hub fromconfigmap "" "group-zones-configmap" (printf "%s-cluster-log-fwd-outputs" (index .ManagedClusterLabels "group-du-sno-zone")) | toLiteral hub}}'
                pipelines: '{{hub fromconfigmap "" "group-zones-configmap" (printf "%s-cluster-log-fwd-pipelines" (index .ManagedClusterLabels "group-du-sno-zone")) | toLiteral hub}}'
        - path: source-crs/PerformanceProfile-MCP-master.yaml
          patches:
            - metadata:
                name: openshift-node-performance-profile
              spec:
                additionalKernelArgs:
                    - rcupdate.rcu_normal_after_boot=0
                    - vfio_pci.enable_sriov=1
                    - vfio_pci.disable_idle_d3=1
                    - efi=runtime
                cpu:
                    isolated: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-cpu-isolated" (index .ManagedClusterLabels "hardware-type")) hub}}'
                    reserved: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-cpu-reserved" (index .ManagedClusterLabels "hardware-type")) hub}}'
                hugepages:
                    defaultHugepagesSize: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-hugepages-default" (index .ManagedClusterLabels "hardware-type")) hub}}'
                    pages:
                        - count: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-hugepages-count" (index .ManagedClusterLabels "hardware-type")) | toInt hub}}'
                          size: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-hugepages-size" (index .ManagedClusterLabels "hardware-type")) hub}}'
                realTimeKernel:
                    enabled: true
    - name: group-du-sno-pgt-group-du-sno-sriov-policy
      policyAnnotations:
        ran.openshift.io/ztp-deploy-wave: "100"
      manifests:
        - path: source-crs/SriovNetwork.yaml
          patches:
            - metadata:
                name: sriov-nw-du-fh
              spec:
                resourceName: du_fh
                vlan: '{{hub fromconfigmap "" "site-data-configmap" (printf "%s-sriov-network-vlan-1" .ManagedClusterName) | toInt hub}}'
        - path: source-crs/SriovNetworkNodePolicy-MCP-master.yaml
          patches:
            - metadata:
                name: sriov-nnp-du-fh
              spec:
                deviceType: netdevice
                isRdma: false
                nicSelector:
                    pfNames: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-sriov-node-policy-pfNames-1" (index .ManagedClusterLabels "hardware-type")) | toLiteral hub}}'
                numVfs: 8
                priority: 10
                resourceName: du_fh
        - path: source-crs/SriovNetwork.yaml
          patches:
            - metadata:
                name: sriov-nw-du-mh
              spec:
                resourceName: du_mh
                vlan: '{{hub fromconfigmap "" "site-data-configmap" (printf "%s-sriov-network-vlan-2" .ManagedClusterName) | toInt hub}}'
        - path: source-crs/SriovNetworkNodePolicy-MCP-master.yaml
          patches:
            - metadata:
                name: sriov-nw-du-fh
              spec:
                deviceType: netdevice
                isRdma: false
                nicSelector:
                    pfNames: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-sriov-node-policy-pfNames-2" (index .ManagedClusterLabels "hardware-type")) | toLiteral hub}}'
                numVfs: 8
                priority: 10
                resourceName: du_fh

    2. Create a group PolicyGenTemplate CR. This example PolicyGenTemplate CR configures logging, VLAN IDs, NICs and Performance Profile for the clusters that match the labels listed under spec.bindingRules:

      apiVersion: ran.openshift.io/v1
kind: PolicyGenTemplate
metadata:
  name: group-du-sno-pgt
  namespace: ztp-group
spec:
  bindingRules:
    # These policies will correspond to all clusters with these labels
    group-du-sno-zone: "zone-1"
    hardware-type: "hardware-type-1"
  mcp: "master"
  sourceFiles:
    - fileName: ClusterLogForwarder.yaml # wave 10
      policyName: "group-du-sno-cfg-policy"
      spec:
        outputs: '{{hub fromconfigmap "" "group-zones-configmap" (printf "%s-cluster-log-fwd-outputs" (index .ManagedClusterLabels "group-du-sno-zone")) | toLiteral hub}}'
        pipelines: '{{hub fromconfigmap "" "group-zones-configmap" (printf "%s-cluster-log-fwd-pipelines" (index .ManagedClusterLabels "group-du-sno-zone")) | toLiteral hub}}'

    - fileName: PerformanceProfile.yaml # wave 10
      policyName: "group-du-sno-cfg-policy"
      metadata:
        name: openshift-node-performance-profile
      spec:
        additionalKernelArgs:
        - rcupdate.rcu_normal_after_boot=0
        - vfio_pci.enable_sriov=1
        - vfio_pci.disable_idle_d3=1
        - efi=runtime
        cpu:
          isolated: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-cpu-isolated" (index .ManagedClusterLabels "hardware-type")) hub}}'
          reserved: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-cpu-reserved" (index .ManagedClusterLabels "hardware-type")) hub}}'
        hugepages:
          defaultHugepagesSize: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-hugepages-default" (index .ManagedClusterLabels "hardware-type")) hub}}'
          pages:
            - size: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-hugepages-size" (index .ManagedClusterLabels "hardware-type")) hub}}'
              count: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-hugepages-count" (index .ManagedClusterLabels "hardware-type")) | toInt hub}}'
        realTimeKernel:
          enabled: true

    - fileName: SriovNetwork.yaml # wave 100
      policyName: "group-du-sno-sriov-policy"
      metadata:
        name: sriov-nw-du-fh
      spec:
        resourceName: du_fh
        vlan: '{{hub fromconfigmap "" "site-data-configmap" (printf "%s-sriov-network-vlan-1" .ManagedClusterName) | toInt hub}}'

    - fileName: SriovNetworkNodePolicy.yaml # wave 100
      policyName: "group-du-sno-sriov-policy"
      metadata:
        name: sriov-nnp-du-fh
      spec:
        deviceType: netdevice
        isRdma: false
        nicSelector:
          pfNames: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-sriov-node-policy-pfNames-1" (index .ManagedClusterLabels "hardware-type")) | toLiteral hub}}'
        numVfs: 8
        priority: 10
        resourceName: du_fh

    - fileName: SriovNetwork.yaml # wave 100
      policyName: "group-du-sno-sriov-policy"
      metadata:
        name: sriov-nw-du-mh
      spec:
        resourceName: du_mh
        vlan: '{{hub fromconfigmap "" "site-data-configmap" (printf "%s-sriov-network-vlan-2" .ManagedClusterName) | toInt hub}}'

    - fileName: SriovNetworkNodePolicy.yaml # wave 100
      policyName: "group-du-sno-sriov-policy"
      metadata:
        name: sriov-nw-du-fh
      spec:
        deviceType: netdevice
        isRdma: false
        nicSelector:
          pfNames: '{{hub fromconfigmap "" "group-hardware-types-configmap" (printf "%s-sriov-node-policy-pfNames-2" (index .ManagedClusterLabels "hardware-type")) | toLiteral hub}}'
        numVfs: 8
        priority: 10
        resourceName: du_fh

    To retrieve site-specific configuration values, use the .ManagedClusterName field. This is a template context value set to the name of the target managed cluster.

    To retrieve group-specific configuration, use the .ManagedClusterLabels field. This is a template context value set to the value of the managed cluster’s labels.

  5. Commit the site PolicyGenerator or PolicyGentemplate CR in Git and push to the Git repository that is monitored by the ArgoCD application.

    Subsequent changes to the referenced configmap CR are not automatically synced to the applied policies. You need to manually sync the new configmap changes to update existing PolicyGenerator CRs. See "Syncing new configmap changes to existing PolicyGenerator or PolicyGenTemplate CRs".

    You can use the same PolicyGenerator or PolicyGentemplate CR for multiple clusters. If there is a configuration change, then the only modifications you need to make are to the configmap objects that hold the configuration for each cluster and the labels of the managed clusters.

Syncing new configmap changes to existing PolicyGenerator or PolicyGentemplate CRs

Prerequisites

  • You have installed the OpenShift CLI (oc).

  • You have logged in to the hub cluster as a user with cluster-admin privileges.

  • You have created a PolicyGenerator or PolicyGentemplate CR that pulls information from a configmap CR using hub cluster templates.

Procedure

  1. Update the contents of your configmap CR, and apply the changes in the hub cluster.

  2. To sync the contents of the updated configmap CR to the deployed policy, do either of the following:

    1. Option 1: Delete the existing policy. ArgoCD uses the PolicyGenerator or PolicyGentemplate CR to immediately recreate the deleted policy. For example, run the following command:

      $ oc delete policy <policy_name> -n <policy_namespace>

    2. Option 2: Apply a special annotation policy.open-cluster-management.io/trigger-update to the policy with a different value every time when you update the configmap. For example:

      $ oc annotate policy <policy_name> -n <policy_namespace> policy.open-cluster-management.io/trigger-update="1"

      You must apply the updated policy for the changes to take effect. For more information, see Special annotation for reprocessing.

  3. Optional: If it exists, delete the ClusterGroupUpdate CR that contains the policy. For example:

    $ oc delete clustergroupupgrade <cgu_name> -n <cgu_namespace>

    1. Create a new ClusterGroupUpdate CR that includes the policy to apply with the updated configmap changes. For example, add the following YAML to the file cgr-example.yaml:

      apiVersion: ran.openshift.io/v1alpha1
kind: ClusterGroupUpgrade
metadata:
  name: <cgr_name>
  namespace: <policy_namespace>
spec:
  managedPolicies:
    - <managed_policy>
  enable: true
  clusters:
  - <managed_cluster_1>
  - <managed_cluster_2>
  remediationStrategy:
    maxConcurrency: 2
    timeout: 240

    2. Apply the updated policy:

      $ oc apply -f cgr-example.yaml