Rancher Kubernetes cluster doesn't detect the node failure

[rooty123]

When I shutdown host, in UI it becomes in state Reconnecting, but kubectl shows
NAME STATUS AGE
rancherkub1 Ready 1d
rancherkub2 Ready 1d

when rancherkub2 is down. As a result kubernetes doesn't reschedule containers that were running on this host.

---

Useful	Info
Versions	Rancher v1.1.1 Cattle: v0.165.7 UI: v1.1.12
Access	localauth admin
Orchestration	Kubernetes
Route	hosts.index

[deniseschannon]

I'm unable to reproduce on v1.1.2.

> kubectl get nodes
NAME                 STATUS     AGE
denise-k8stesting1   NotReady   10h
denise-k8stesting2   Ready      10h
denise-k8stesting3   Ready      10h

Can you go to the System stack in Kubernetes and share what image tag the kubernetes service is running? This would help me know what template you are running on kubernetes.

[rooty123]

I also was able to reproduce on 2.0-pre1. Here is screenshot:

Perhaps you need to test with 2 nodes (not 3)

[deniseschannon]

@rooty123 How did you get Kubernetes to be running v1.3? This version of Kubernetes is not supported in Rancher v1.1.x series.

I can tell it's running v1.3 with the image being rancher/k8s:v1.3.0-rancher3

[rooty123]

@deniseschannon I just use latest 1.2.0-pre release, which supports upgrade to 1.3

[aemneina]

hey @rooty123 can you check with a more modern version of rancher server? If you hit it again, please do reopen or file a new issue.