openshift-hyperkube RPM is too big

It was notices that the openshift-hyperkube RPM which is primarilly, perhaps exclusively, used to install the kubelet in RHCOS or other environments included kube-apiserver, kube-controller-manager, and kube-scheduler binaries. Those binaries are all built and used via container images, which as far as I can tell don't make use of the RPM.    

4.12 - 4.16

100%

    1. rpm -ql openshift-hyperkube on any node
    2.
    3.
    

# rpm -ql openshift-hyperkube
/usr/bin/hyperkube
/usr/bin/kube-apiserver
/usr/bin/kube-controller-manager
/usr/bin/kube-scheduler
/usr/bin/kubelet
/usr/bin/kubensenter

# ls -lah /usr/bin/kube-apiserver /usr/bin/kube-controller-manager /usr/bin/kube-scheduler /usr/bin/hyperkube /usr/bin/kubensenter /usr/bin/kubelet
-rwxr-xr-x. 2 root root  945 Jan  1  1970 /usr/bin/hyperkube
-rwxr-xr-x. 2 root root 129M Jan  1  1970 /usr/bin/kube-apiserver
-rwxr-xr-x. 2 root root 114M Jan  1  1970 /usr/bin/kube-controller-manager
-rwxr-xr-x. 2 root root  54M Jan  1  1970 /usr/bin/kube-scheduler
-rwxr-xr-x. 2 root root 105M Jan  1  1970 /usr/bin/kubelet
-rwxr-xr-x. 2 root root 3.5K Jan  1  1970 /usr/bin/kubensenter

Just the kubelet and deps on the host OS, that's all that's necessary

My proposed change would be for people that cared about making this slim to install `openshift-hyperkube-kubelet` instead.