KubeDNS : Headless Service - Records are overwrited when using deployment replicas

[gunboe]

/area dns

What happened:

KubeDNS overwrite the last replica records (IP/Host) in subdomain, instead of appending the records in KubeDNS.

What you expected to happen:

I'm trying to make a DNS "loadbalance" with Headless Service (https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/) with replicas from a deployment.
Trying to resolve a $hostname.$subdomain.$namespace.svc.cluster.local from KubeDNS, I expect to receive a IP list to the same $hostname.

Address 1: 192.168.103.153 hostXYZ.subdomain.nspace.svc.cluster.local
Address 3: 192.168.62.195 hostXYZ.subdomain.nspace.svc.cluster.local
Address 4: 192.168.80.144  hostXYZ.subdomain.nspace.svc.cluster.local

How to reproduce it (as minimally and precisely as possible):

Here you are Yaml file:

kind: Deployment
metadata:
  labels:
    run: busy2
    name: busybox
  name: busy2
  namespace: gsopi-p
spec:
  replicas: 3
  selector:
    matchLabels:
      run: busy2
  template:
    metadata:
      creationTimestamp: null
      labels:
        run: busy2
        name: busybox
    spec:
      hostname: busy2
      subdomain: default-subdomain
      containers:
      - args:
        - sh
        image: busybox
        imagePullPolicy: Always
        name: busy2
        resources: {}
        stdin: true
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        tty: true
      dnsPolicy: ClusterFirst
      restartPolicy: Always
---
apiVersion: v1
kind: Service
metadata:
  name: default-subdomain
spec:
  selector:
    name: busybox
  clusterIP: None
  ports:
  - name: foo # Actually, no port is needed.
    port: 80
    targetPort: 80

Testing:

$ kubectl get pods -o wide  
NAME                                 READY     STATUS    RESTARTS   AGE       IP               NODE
busy1-4067030300-vb0vj               1/1       Running   0          4d        192.168.51.172   kn00006
busy2-4281529631-5j3rv               1/1       Running   0          2h        192.168.90.231   kn00005
busy2-4281529631-df66c               1/1       Running   0          2h        192.168.80.146   kn00004
busy2-4281529631-jc4rg               1/1       Running   0          2h        192.168.62.14    kn00007

$ kubectl exec -it busy1-4067030300-vb0vj sh  
/ # nslookup default-subdomain.gsopi-p.svc.cluster.local
Server:    10.36.192.10
Address 1: 10.36.192.10 kube-dns.kube-system.svc.cluster.local

Name:      default-subdomain.gsopi-p.svc.cluster.local
Address 1: 192.168.90.231 busy2.default-subdomain.gsopi-p.svc.cluster.local

Environment:

Client Version: version.Info{Major:"1", Minor:"6", GitVersion:"v1.6.8", GitCommit:"d74e09bb4e4e7026f45becbed8310665ddcb8514", GitTreeState:"clean", BuildDate:"2017-08-03T18:12:08Z", GoVersion:"go1.7.6", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"6", GitVersion:"v1.6.8+coreos.0", GitCommit:"fc34f797fe56c4ab78bdacc29f89a33ad8662f8c", GitTreeState:"clean", BuildDate:"2017-08-05T00:01:34Z", GoVersion:"go1.7.6", Compiler:"gc", Platform:"linux/amd64"}```

[k8s-github-robot]

@gunboe
There are no sig labels on this issue. Please add a sig label by:

1. mentioning a sig: @kubernetes/sig-<group-name>-<group-suffix>
   e.g., @kubernetes/sig-contributor-experience-<group-suffix> to notify the contributor experience sig, OR

2. specifying the label manually: /sig <label>
   e.g., /sig scalability to apply the sig/scalability label

Note: Method 1 will trigger an email to the group. You can find the group list here and label list here.
The <group-suffix> in the method 1 has to be replaced with one of these: bugs, feature-requests, pr-reviews, test-failures, proposals

[guillaumerose]

This issue is due to the dns. See kubernetes/dns#116.

[gunboe]

So, will be nice to be implemented...
/kind feature

[dims]

we don't need to track it in both repos. let's close this one since we have 116 in kubernetes/dns repo.

/assign

[dims]

/close