Env: Vbox on a windows 10 desktop machine Two ubuntu VMs, one VM is master and the other one is k8s(1.7) worker.

I can see two nodes are "ready" when get nodes. But even deploy a very simple nginx pod, I got the error message from pod describe "norm | SandboxChanged |Pod sandbox changed, it will be killed and re-created." and "warning | FailedSync| Error syncing pod".

But if I run the docker container directly on the worker, the container can be up and running. Anyone has a suggestion what I can check for?

k8s-master@k8smaster-VirtualBox:~$ **kubectl get pods** NAME          
READY     STATUS             RESTARTS   AGE
movie-server-1517284798-lbb01   0/1       CrashLoopBackOff   6        
16m

k8s-master@k8smaster-VirtualBox:~$ **kubectl describe pod
movie-server-1517284798-lbb01**
--- clip ---  kubelet, master-virtualbox       spec.containers{movie-server}   Warning FailedError: failed to start
container "movie-server": Error response from daemon:
{"message":"cannot join network of a non running container:
3f59947dbd404ecf2f6dd0b65dd9dad8b25bf0c418aceb8cf666ad0761402b53"}

kubelet, master-virtualbox       spec.containers{movie-server}  
Warning BackOffBack-off restarting failed container

kubelet, master-virtualbox                                      Normal
SandboxChanged          Pod sandbox changed, it will be killed and
re-created.

kubelet, master-virtualbox      spec.containers{movie-server}   Normal
PulledContainer image "nancyfeng/movie-server:0.1.0" already present
on machine

kubelet, master-virtualbox       spec.containers{movie-server}  
Normal  CreatedCreated container

kubelet, master-virtualbox                                      
Warning FailedSync              Error syncing pod

kubelet, master-virtualbox       spec.containers{movie-server}  
Warning FailedError: failed to start container "movie-server": Error
response from daemon: {"message":"cannot join network of a non running
container:
72ba77b25b6a3969e8921214f0ca73ffaab4c82d8a2852e3d1b1f3ac5dde6ce1"}
--- clip ---