Fix operator to start arktos gRPC server only for Arktos env. Optionally load config using KUBECONFIG path if specified.

[vinaykul]

What type of PR is this?

Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line:

/kind design /kind feature /kind bug /kind cleanup /kind documentation

What this PR does / why we need it: This allows using a proxy to route node vs all other requests to API and node info aggregator for arktos scaleout usage.

Which issue(s) this PR fixes:

Fixes #

Special notes for your reviewer:

• Verified using kind setup that default case works.
• Verified that proxy works by doing the following:
  • Start 2-node k8s cluster using kubeadm
  • Start proxy (kubectl proxy -p 8080
  • Create /etc/kubernetes/mzadmin.conf as follows:

    
    root@ip-192-168-1-144:~# cat /etc/kubernetes/proxy_admin.conf 
    apiVersion: v1
    clusters:

• cluster: server: http://127.0.0.1:8080 name: kubernetes contexts:
• context: cluster: kubernetes user: kubernetes-admin name: kubernetes-admin@kubernetes current-context: kubernetes-admin@kubernetes root@ip-192-168-1-144:~#

  - Deploy mizar as follows:

  root@ip-192-168-1-144:~# cat proxy_deploy_vmizar.yaml ... containers:

  • name: mizar-operator image: skiibum/endpointopr:dev env:
    • name: KUBECONFIG value: "/admin.conf"
    • name: FEATUREGATE_BWQOS value: 'false' securityContext: privileged: true volumeMounts:
    • name: kubeconfig mountPath: /admin.conf volumes:
  • name: kubeconfig hostPath: path: /etc/kubernetes/proxy_admin.conf type: File

Does this PR introduce a user-facing change?: No

[vinaykul]

LGTM assuming it's e2e tested in kind environment.

Yes. Deployed netpods and ping tested in kind env. CI should have covered the real cluster case and unit tests.