Uploaded image for project: 'OpenShift Virtualization'
  1. OpenShift Virtualization
  2. CNV-73754

Show details on how to determine the best defaultCPUModel for the Cluster

XMLWordPrintable

    • Icon: Feature Request Feature Request
    • Resolution: Unresolved
    • Icon: Critical Critical
    • None
    • None
    • None
    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • False
    • None

      On clusters with heterogeneus CPUs, the customer needs to use the process described in 9.15.4.1. Configuring the default CPU model

      While the procedure shows how to set the CPU model, is doesn't have any advice or steps on how to determine the optimal CPU model.

      The reader may go about the wrong way or set an unsuitable model if we do not provide the general guidelines and directions.

      Requirements:

      1. The named CPU model be the highest common denominator for all nodes (not higher than any single node). Use the labels on the nodes, then find the highest common model for all of them.
      2. For VMs that needs special features only available on certain nodes, recommend the user to specify in the VM YAML or create a VirtualMachineClusterPreference with enables the feature, and apply it to the VMs. For example:
        # oc get virtualmachineclusterpreference homelab-vm -o yaml | yq '.spec.cpu'
        preferredCPUFeatures:
          - name: vmx
            policy: require 

              rhn-support-mtessun Martin Tessun
              rhn-support-gveitmic Germano Veit Michel
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated: