Values for NodeSelectorTerms are evaluated asynchronously

[JamesGAWS]

Currently for AWS Elastic Kubernetes Serivce it is possible to create a deployment with the following spec:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: inflate
spec:
  replicas: 1
  selector:
    matchLabels:
      app: inflate
  template:
    metadata:
      labels:
        app: inflate
    spec:
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            nodeSelectorTerms:
            - matchExpressions:
              - key: "eks.amazonaws.com/nodegroup"
                operator: In
                values:
                - "${INVALID_VALUE}"
      terminationGracePeriodSeconds: 0
      securityContext:
        runAsUser: 1000
        runAsGroup: 3000
        fsGroup: 2000
      containers:
      - name: inflate
        image: public.ecr.aws/eks-distro/kubernetes/pause:3.7
        resources:
          requests:
            cpu: 1
        securityContext:
          allowPrivilegeEscalation: false

Despite the presence of invalid characters in the nodeSelectorTerms value, the API server will accept this payload. During scheduling, the label value is validated which the label will fail since it contains invalid characters.

Typical behavior for invalid characters is for the apply call to fail synchronously. Why is this field different?

[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

[k8s-triage-robot]

The Kubernetes project currently lacks enough active 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 rotten
• Close this issue with /close
• Offer to help out with Issue Triage

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

/lifecycle rotten