kubernetes: Containers from init-containers section do not execute during node reboots

BUG REPORT Kubernetes version v1.4.3+coreos.0

Environment:

• Cloud provider or hardware configuration: Bare metal CoreOS installation. 3 management nodes + 2 worker nodes.
• OS (e.g. from /etc/os-release): CoreOS stable (1185.3.0)
• Kernel 4.7.3-coreos-r2

What happened: I added an init-container to the kube-scheduler.yaml file in /etc/kubernetes/manifests directory. It was executed once after the yaml file modification. When I rebooted the node running this POD, the POD was restarted but the init-container did not execute:

The kube-scheduler.yaml file:

apiVersion: v1
kind: Pod
metadata:
  name: kube-scheduler
  namespace: kube-system
  annotations:
    pod.beta.kubernetes.io/init-containers: '[
      {
        "image": "quay.io/coreos/hyperkube:v1.4.3_coreos.0",
        "name": "wait-for-master",
        "command": [ "/bin/bash", "-c", "while ! timeout 1 bash -c \"/kubectl --server=http://127.0.0.1:8080/ cluster-info\"; do sleep 1; done" ]
      }
    ]'
spec:
  hostNetwork: true
  containers:
  - name: kube-scheduler
    image: quay.io/coreos/hyperkube:v1.4.3_coreos.0
    command:
    - /hyperkube
    - scheduler
    - --master=http://127.0.0.1:8080
    - --leader-elect=true
    livenessProbe:
      httpGet:
        host: 127.0.0.1
        path: /healthz
        port: 10251
      initialDelaySeconds: 15
      timeoutSeconds: 1

The POD description after the node reboot:

Name:           kube-scheduler-10.54.147.6
Namespace:      kube-system
Node:           10.54.147.6/10.54.147.6
Start Time:     Wed, 09 Nov 2016 07:51:35 +0100
Labels:         <none>
Status:         Running
IP:             10.54.147.6
Controllers:    <none>
Init Containers:
  wait-for-master:
    Container ID:       docker://ff5ef0e759eb332c50fcc117a5758c41080835addbd76e33eb60cdde10d9587b
    Image:              quay.io/coreos/hyperkube:v1.4.3_coreos.0
    Image ID:           docker://sha256:5d3bc50f81574bc41952e13862b04b98f310ab60601ff990e7267a4ba5227e8b
    Port:
    Command:
      /bin/bash
      -c
      while ! timeout 1 bash -c "/kubectl --server=http://127.0.0.1:8080/ cluster-info"; do sleep 1; done
    State:                      Terminated
      Reason:                   Completed
      Exit Code:                0
      Started:                  Tue, 08 Nov 2016 14:17:27 +0100
      Finished:                 Tue, 08 Nov 2016 14:17:35 +0100
    Ready:                      True
    Restart Count:              0
    Volume Mounts:              <none>
    Environment Variables:      <none>
Containers:
  kube-scheduler:
    Container ID:       docker://03e40bb84d477e6a284c4bfdf677d15135fe78785b17ae62d09cd04063f81aec
    Image:              quay.io/coreos/hyperkube:v1.4.3_coreos.0
    Image ID:           docker://sha256:5d3bc50f81574bc41952e13862b04b98f310ab60601ff990e7267a4ba5227e8b
    Port:
    Command:
      /hyperkube
      scheduler
      --master=http://127.0.0.1:8080
      --leader-elect=true
    State:                      Running
      Started:                  Wed, 09 Nov 2016 07:51:36 +0100
    Last State:                 Terminated
      Reason:                   Error
      Exit Code:                2
      Started:                  Tue, 08 Nov 2016 20:50:42 +0100
      Finished:                 Wed, 09 Nov 2016 07:48:44 +0100
    Ready:                      True
    Restart Count:              4
    Liveness:                   http-get http://127.0.0.1:10251/healthz delay=15s timeout=1s period=10s #success=1 #failure=3
    Volume Mounts:              <none>
    Environment Variables:      <none>
Conditions:
  Type          Status
  Initialized   True 
  Ready         True 
  PodScheduled  True 
No volumes.
QoS Class:      BestEffort
Tolerations:    <none>

How to reproduce it Create a pod with the cloud-init section and pin it to one node. Reboot the node several times and check if the cloud-init container was executed.