[Feature] Leverage Affinity rather than nodeSelector to target Control Plane Nodes

[brownz11]

Is this a BUG REPORT or FEATURE REQUEST?:

Uncomment only one, leave it on its own line:

/kind bug
/kind feature

What happened:
The current manifests use a nodeSelector to target controlplane node. Specifically an empty label. Some distributions may leverage other labels, or populate the label value. This causes the pods to not get scheduled.

What you expected to happen:
The pods to be scheduled to control plane nodes.

How to reproduce it (as minimally and precisely as possible):
Label controlplane nodes with "node-role.kubernetes.io/control-plane=true" rather than empty.

Anything else we need to know?:
An easy fix, and what I've seen in other projects, as well as tested on my clusters is to remove the nodeSelector and instead add a nodeAffinity that matches the label rather than the label AND it's contents.

nodeAffinity:
    requiredDuringSchedulingIgnoredDuringExecution:
      nodeSelectorTerms:
        - matchExpressions:
            - key: node-role.kubernetes.io/control-plane
              operator: Exists
        - matchExpressions:
            - key: node-role.kubernetes.io/controlplane
              operator: Exists
        - matchExpressions:
            - key: node-role.kubernetes.io/master
              operator: Exists

Environment:

• csi-vsphere version: 3.1.1
• vsphere-cloud-controller-manager version: N/a
• Kubernetes version: 1.25
• vSphere version: N/a
• OS (e.g. from /etc/os-release): N/a
• Kernel (e.g. uname -a): N/a
• Install tools: N/a
• Others: N/a

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[vdkotkar]

/assign vdkotkar

[chethanv28]

• matchExpressions:
  - key: node-role.kubernetes.io/controlplane
  operator: Exists

This looks like additional selector with a typo. It should have been control-plane and it is already part of the selector above.

[brownz11]

It is not a typo, at least not mine. Some k8s distributions use "control-plane" other use "controlplane".

The Azure CSI driver already uses the affinity stanza provided above.

https://github.com/kubernetes-sigs/azuredisk-csi-driver/blob/bc6ce67e47376ccea4c639dadc7fb5e3feb17ad5/deploy/v1.30.0/csi-snapshot-controller.yaml#L38

And there's a lot of other projects that use the affinity explicitly to avoid this "control-plane" vs "controlplane" issue.