kubernetes: Intermittent error when doing kubectl apply -f for a HorizontalPodAutoscaler: 'unable to find api field in struct HorizontalPodAutoscalerSpec for the json field "scaleTargetRef"'

What keywords did you search in Kubernetes issues before filing this one?

• unable to find api field in struct HorizontalPodAutoscalerSpec for the json field “scaleTargetRef”
• HorizontalPodAutoscaler
• scaleTargetRef

Also searched google/stackoverflow

---

Is this a BUG REPORT or FEATURE REQUEST? Bug report

Kubernetes version (use kubectl version): v1.4.0

Environment:

• GKE with kubernetes v1.4.0

What happened: Do kubectl apply -f my-service.yaml

Intermittently, since upgrading to 1.4.0 I have been getting

error: error when applying patch:

to:
&{0xc8203546c0 0xc8202e3340 default my-service-horizpodautoscaler /my-service/build/k8s/my-service.yaml &HorizontalPodAutoscaler{ObjectMeta:k8s_io_kubernetes_pkg_api_v1.ObjectMeta{Name:my-service-horizpodautoscaler,GenerateName:,Namespace:,SelfLink:,UID:,ResourceVersion:,Generation:0,CreationTimestamp:0001-01-01 00:00:00 +0000 UTC,DeletionTimestamp:<nil>,DeletionGracePeriodSeconds:nil,Labels:map[string]string{},Annotations:map[string]string{kubectl.kubernetes.io/last-applied-configuration: ,},OwnerReferences:[],Finalizers:[],ClusterName:,},Spec:HorizontalPodAutoscalerSpec{ScaleTargetRef:CrossVersionObjectReference{Kind:Deployment,Name:my-service-frontend,APIVersion:extensions/v1beta1,},MinReplicas:*1,MaxReplicas:3,TargetCPUUtilizationPercentage:nil,},Status:HorizontalPodAutoscalerStatus{ObservedGeneration:nil,LastScaleTime:<nil>,CurrentReplicas:0,DesiredReplicas:0,CurrentCPUUtilizationPercentage:nil,},} &TypeMeta{Kind:,APIVersion:,} 1537793 false}
for: "/my-service/build/k8s/my-service.yaml": error when creating patch with:
original:
{"kind":"HorizontalPodAutoscaler","apiVersion":"autoscaling/v1","metadata":{"name":"my-service-horizpodautoscaler","creationTimestamp":null},"spec":{"scaleTargetRef":{"kind":"Deployment","name":"my-service-frontend","apiVersion":"extensions/v1beta1"},"minReplicas":1,"maxReplicas":3},"status":{"currentReplicas":0,"desiredReplicas":0}}
modified:
{"kind":"HorizontalPodAutoscaler","apiVersion":"autoscaling/v1","metadata":{"name":"my-service-horizpodautoscaler","creationTimestamp":null,"annotations":{"kubectl.kubernetes.io/last-applied-configuration":"{\"kind\":\"HorizontalPodAutoscaler\",\"apiVersion\":\"autoscaling/v1\",\"metadata\":{\"name\":\"my-service-horizpodautoscaler\",\"creationTimestamp\":null},\"spec\":{\"scaleTargetRef\":{\"kind\":\"Deployment\",\"name\":\"my-service-frontend\",\"apiVersion\":\"extensions/v1beta1\"},\"minReplicas\":1,\"maxReplicas\":3},\"status\":{\"currentReplicas\":0,\"desiredReplicas\":0}}"}},"spec":{"scaleTargetRef":{"kind":"Deployment","name":"my-service-frontend","apiVersion":"extensions/v1beta1"},"minReplicas":1,"maxReplicas":3},"status":{"currentReplicas":0,"desiredReplicas":0}}
current:
{"kind":"HorizontalPodAutoscaler","apiVersion":"extensions/v1beta1","metadata":{"name":"my-service-horizpodautoscaler","namespace":"default","selfLink":"/apis/autoscaling/v1/namespaces/default/horizontalpodautoscalers/my-service-horizpodautoscaler","uid":"REDACTED-REDA-REDA-REDA-REDACTED","resourceVersion":"1537793","creationTimestamp":"2016-09-28T15:51:27Z","annotations":{"kubectl.kubernetes.io/last-applied-configuration":"{\"kind\":\"HorizontalPodAutoscaler\",\"apiVersion\":\"autoscaling/v1\",\"metadata\":{\"name\":\"my-service-horizpodautoscaler\",\"creationTimestamp\":null},\"spec\":{\"scaleTargetRef\":{\"kind\":\"Deployment\",\"name\":\"my-service-frontend\",\"apiVersion\":\"extensions/v1beta1\"},\"minReplicas\":1,\"maxReplicas\":3},\"status\":{\"currentReplicas\":0,\"desiredReplicas\":0}}"}},"spec":{"scaleRef":{"kind":"Deployment","name":"my-service-frontend","apiVersion":"extensions/v1beta1","subresource":"scale"},"minReplicas":1,"maxReplicas":3,"cpuUtilization":{"targetPercentage":80}},"status":{"lastScaleTime":"2016-10-08T23:19:12Z","currentReplicas":1,"desiredReplicas":1,"currentCPUUtilizationPercentage":0}}

for: "/my-service/build/k8s/my-service.yaml": unable to find api field in struct HorizontalPodAutoscalerSpec for the json field "scaleTargetRef"

As you can see, it seems that the “current” metadata specifies “apiVersion” as extensions/v1beta1, that might be correct, but it seems suspicious to me…

I have the feeling it works every second time I do apply, but hard to say from my small sample size.

What you expected to happen: The apply to work every time.

How to reproduce it

I used this yaml file (cleaned)

apiVersion: v1
kind: Service
metadata:
  name: my-service
  labels:
    app: my-service
    tier: frontend
spec:
  type: NodePort
  ports:
  - name: http8080
    protocol: TCP
    port: 8080
    targetPort: 8080
  selector:
    app: my-service
    tier: frontend

---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: my-service-frontend
spec:
  replicas: 1
  template:
    metadata:
      labels:
        app: my-service
        tier: frontend
    spec:
      containers:
      - name: my-service
        image: eu.gcr.io/my-project/my-service:v0.0.0
        env:
        - name: MY_ENVIRONMENT
          value: dev
        ports:
        - containerPort: 8080
        resources:
          requests:
            cpu: 100m
            memory: 800Mi

---
apiVersion: autoscaling/v1 #I've tried with extensions/v1beta1 and "scaleRef" below, but that doesn't work properly either (similar error message)
kind: HorizontalPodAutoscaler
metadata:
  name: my-service-horizpodautoscaler
spec:
  scaleTargetRef:
    kind: Deployment
    name: my-service-frontend
    apiVersion: extensions/v1beta1 #I've tried with and without this, makes no difference
  minReplicas: 1
  maxReplicas: 3

Anything else do we need to know:

GET on clusters using https://cloud.google.com/container-engine/reference/rest/v1/projects.zones.clusters/get reveals:

[...]
 "addonsConfig": {
  "httpLoadBalancing": {
  },
  "horizontalPodAutoscaling": {
  }
 },
[...]
   "initialNodeCount": 1,
   "autoscaling": {
    "enabled": true,
    "minNodeCount": 1,
    "maxNodeCount": 3
   },
[...]

 "initialClusterVersion": "1.3.6",
 "currentMasterVersion": "1.4.0",
 "currentNodeVersion": "1.4.0",

To solve this I have temporarily moved the HPA section out of the file into a separate file, but I prefer having the complete configuration in one place.

Am I doing something wrong?

Thanks!