Kubernetes not finding nodes

[ScyDev GmbH]

Sometimes when I delete a replication controller and create a new one, everything stops working until I re-create the whole cluster.
First I tried just deleting the node and let the cluster create a new one automatically, but that doesn't help always.
Right now I have a node running, but Kubernetes can't see it, even after multiple tries.

What could be the cause of this?


$ gcloud compute instances list
NAME                                                ZONE           MACHINE_TYPE PREEMPTIBLE INTERNAL_IP EXTERNAL_IP    STATUS
gke-test-cluster-default-pool-0bd8d63e-pmmt europe-west1-c n1-highcpu-2             10.132.0.2  130.211.71.107 RUNNING


$ kubectl get nodes
$


$ kubectl get events -w
FIRSTSEEN                        LASTSEEN                         COUNT     NAME                     KIND      SUBOBJECT   TYPE      REASON             SOURCE                 MESSAGE
2016-05-07 20:13:02 +0200 CEST   2016-05-07 20:13:17 +0200 CEST   5         mongo-controller-4oz5x   Pod                   Warning   FailedScheduling   {default-scheduler }   no nodes available to schedule pods


Thanks in advance.

- Lukas

[George (Google Cloud Support)]

Hello Lukas,

This gce-discussion group is dedicated for GCE related discussions. However, you can post your question on the proper group or on Stackoverflow with the appropriate tags and make sure to include all the relevant information as Google's engineers are active there.

However you can run the following command for more information about the node:

kubectl describe nodes

You can check /var/log/kubelet.log for some information about the node.

I hope this helps.

Sincerely,

George

[ScyDev GmbH]

Hi George

I'm sorry, I assumed this was the right place since I saw GKE as part of GCE.
Thanks for redirecting me and giving me a hint.

Best regards,
Lukas