What keywords did you search in Kubernetes issues before filing this one? (If you have found any duplicates, you should instead reply there.): Pod status started null, started, 01 Jan 0001

Kubernetes version (use kubectl version):

Client Version: version.Info{Major:"1", Minor:"6", GitVersion:"v1.6.3", GitCommit:"0480917b552be33e2dba47386e51decb1a211df6", GitTreeState:"clean", BuildDate:"2017-05-10T15:48:59Z", GoVersion:"go1.7.5", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"6", GitVersion:"v1.6.3", GitCommit:"0480917b552be33e2dba47386e51decb1a211df6", GitTreeState:"clean", BuildDate:"2017-05-10T15:38:08Z", GoVersion:"go1.7.5", Compiler:"gc", Platform:"linux/amd64"}`

Environment:

• Cloud provider or hardware configuration: -
• OS (e.g. from /etc/os-release): CentOS 7
• Kernel (e.g. uname -a): Linux 3.10.0-514.16.1.el7.x86_64 #1 SMP Wed Apr 12 15:04:24 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
• Install tools: kubeadm 1.6

What happened:
When a pod is created and starts running, kubectl describe pod-name shows the correct container start time in Containers -> containername -> State -> Started. After the container is terminated, the start time changes to Mon, 01 Jan 0001 00:00:00 +0000. In addition, kubectl get pod-name -o json shows that status.containerStatuses[0].state.terminated.startedAt is null.

What you expected to happen:
I'm expecting the start time to not change after the pod is terminated, so that afterwards I could check how long a time the container execution took.

How to reproduce it (as minimally and precisely as possible):
Have a running kubernetes cluster, and create a pod with the following command:

kubectl create -f - <<EOF
apiVersion: v1
kind: Pod
metadata:
  name: sleep
spec:
  containers:
  - name: sleep
    image: ubuntu
    command: ["sleep"]
    args: ["30"]
  restartPolicy: Never
EOF

This created a container that sleeps for 30 seconds and quits. During those 30 seconds, check the pod status with kubectl describe po sleep:

...
State:		Running
  Started:      Wed, 17 May 2017 10:25:22 +0300
...

After 30 seconds, check the status of the terminated container with kubectl describe po sleep again:

...
State:		Terminated
  Reason:	Completed
  Exit Code:	0
  Started:	Mon, 01 Jan 0001 00:00:00 +0000
  Finished:	Wed, 17 May 2017 10:25:52 +0300
...

With kubectl get po sleep -o json we see that the container's start time is null:

...
"state": {
    "terminated": {
        "containerID": "docker://ca0dbe0dab3bb91ad0b05ac167e44d5fbf805f736ef98f61b9e106ba4456667f",
        "exitCode": 0,
        "finishedAt": "2017-05-17T07:25:52Z",
        "reason": "Completed",
        "startedAt": null
    }
}
...

Anything else we need to know:
I need to measure container run durations and this bug prevents me from doing so.