Welcome to the Linux Foundation Forum!

CH 13: Exercise 13.3: Adding tools for monitoring and metrics

Hi All,

I followed the steps the lab everything seem to work.
When I try to connect to my node port from my host machine

Host Linux Mint 20.3
VMWare Workstation Pro 16

https://192,168.100.147:30444

{
"kind": "Status",
"apiVersion": "v1",
"metadata": {},
"status": "Failure",
"message": "forbidden: User \"system:anonymous\" cannot get path \"/\"",
"reason": "Forbidden",
"details": {},
"code": 403
}

I will upload the output of several commands

Comments

  • chrispokorni
    chrispokorni Posts: 2,155

    Hi @bdkdavid,

    Thank you for all the detailed outputs attached.

    The first concern is about the node IP addresses and the pod IP addresses. If nodes are on the 192.168.100.0/x network, what IP addresses are used by calico for pods? I see pods on 192.168.69.a, 192,168.76.b, and 192.168.114.c subnets. I suspect that the pod IP addresses overlap the node IP addresses - which should be avoided in a cluster.

    The second concern is about the helm chart used to deploy the kubernetes-dashboard release. The name of the chart is metrics-server-3.8.2 (which deploys a metrics-server, not a kubernetes-dashboard). The chart should have been kubernetes-dashboard-.... for a kubernetes-dashboard installation. Here I'd recommend uninstalling with helm the current kubernetes-dashboard release, and use helm to install the kubernetes-dashboard from a kubernetes-dashboard chart.

    The kubernetes-dashboard chart and general installation instructions can be found at:

    https://artifacthub.io/packages/helm/k8s-dashboard/kubernetes-dashboard

    Regards,
    -Chris

Categories

Upcoming Training