Cluster explorer and cluster k8s shell not working


I have a rancher 2.5.9 instance running from docker on one subnet, and a rancher created (nodepool) vsphere cluster running in another subnet created by this rancher instance.

Most things work fine. Regular cluster manager works, I can create new k8s resources from the UI, I can see status of nodes, there’s no errors showing on the cluster main page in the original UI.

I have two big problems:

  1. Whenever I try to open Cluster Explorer for any of my clusters except the local cluster, I get that giant HTTP 500 page with the pretty artwork that says " HTTP Error 500: Internal Server Error from /k8s/clusters/c-qxrhn/v1/schemas". If I use browser diagnostic tools, I can see a request for that schema resource returning 500, with the text of the response being “invalid upgrade response: status code 200”

  2. I cannot open k8s shell for clusters other than local.

I can confirm that the api is reachable on the control plane nodes at port 6443 from the rancher server. The cattle agent seems to be running, and I can curl the rancher server URL on port 443 from inside the cattle agent container via k8s exec.

One possible clue is that on the local cluster, I see a bunch of failed pods for helm-operation with the image being rancher/shell:XX with logs eventually showing " Waiting for Kubernetes API to be available Timeout waiting for kubernetes". I am not sure what to troubleshoot here, but clearly some kind of connectivity isn’t happening. Do I need some port to the destination cluster nodes other than port 6443?


My problem was exactly the same.
The solution was a wrong configured firewall.
Sometimes, it can be so easy :slightly_smiling_face:

Poderia dar mais detalhes de como solucionou o caso? Qual alteração realizou no firewall? Estou com o mesmo problema relatado pelo @miwood , porém não consegui resolver ainda.