When configuring memory manager in a KubeletConfig, it errors when using lower case "static" policy (like CPUManager uses), ex:

apiVersion: machineconfiguration.openshift.io/v1
kind: KubeletConfig
metadata:
  name: pin
spec:
  machineConfigPoolSelector:
    matchLabels:
      custom-kubelet: cpumanager-enabled
  kubeletConfig:
     cpuManagerPolicy: static
     cpuManagerReconcilePeriod: 5s
     reservedSystemCPUs: "0,1,2,3,112,113,114,115"
     topologyManagerPolicy: single-numa-node
     memoryManagerPolicy: static
     reservedMemory:
       - numaNode: 0
         limits:
           memory: "1124Mi"

Resulted in the node not starting kubelet due to:
Oct 13 16:13:02 worker-perf52 hyperkube[17970]: E1013 16:13:02.747246   17970 container_manager_linux.go:360] "Failed to initialize memory manager" err="unknown policy: \"static\""
Oct 13 16:13:02 worker-perf52 hyperkube[17970]: E1013 16:13:02.747255   17970 server.go:294] "Failed to run kubelet" err="failed to run Kubelet: unknown policy: \"static\""
Oct 13 16:13:02 worker-perf52 systemd[1]: kubelet.service: Main process exited, code=exited, status=1/FAILURE

This was only resolved by changing the policy to upper case "Static"

Ideally it could accept "static" or "Static", but at least it would be less confusing if it matched the example CPUManager lower case "static" syntax, ex: https://docs.openshift.com/container-platform/4.11/scalability_and_performance/using-cpu-manager.html

4.9

Every time "static" is used for memory manager policy the node will not become Ready due to kubelet failure.

1. oc create -f KubeletConfig.yaml (from above)
2. check node journal logs to confirm why kubelet is failing
3. recover by manually editing the policy to be "Static"

Kubelet fails to start

Policy is not case sensitive or at least matches the long used CPUManager lower case

Also OpenShift docs showing an example memory manager KubeletConfig could be very helpful!