Skip to content
dubcluster

Connect to the cluster

You’ll do this once. After that, kubectl just works.

1. Join the tailnet

Section titled “1. Join the tailnet”

The cluster isn’t on the public internet — you reach it over Tailscale, a private mesh network.

  1. Install Tailscale on your machine and sign in.

  2. Accept the invite to the tailnet that you were sent.

  3. Confirm you can see the cluster:

    Terminal window
    tailscale status        # you should be connected
    ping 100.73.243.2       # the cluster API server

2. Install the tools

Section titled “2. Install the tools”

You need two things: kubectl (the Kubernetes CLI) and kubelogin (the plugin that handles the Google login).

macOS (Homebrew):

Terminal window
brew install kubectl int128/kubelogin/kubelogin

Windows (Scoop) or download the binaries:

Terminal window
scoop install kubectl
# kubelogin: download from https://github.com/int128/kubelogin/releases,
# rename the binary to kubectl-oidc_login.exe, and put it on your PATH.

Linux:

Terminal window
# kubectl: https://kubernetes.io/docs/tasks/tools/
# kubelogin via krew:
kubectl krew install oidc-login

3. Drop in your kubeconfig

Section titled “3. Drop in your kubeconfig”

You were sent a kubeconfig file named something like dubcluster-<you>.yaml. Save it:

Terminal window
mkdir -p ~/.kube
mv ~/Downloads/dubcluster-<you>.yaml ~/.kube/dubcluster.yaml

4. Log in

Section titled “4. Log in”

Run any kubectl command against the file. The first time, your browser opens for a Google login — use the Google account you were onboarded with.

Terminal window
kubectl --kubeconfig ~/.kube/dubcluster.yaml get pods

Success looks like:

No resources found in tenant-<you> namespace.

That empty result is correct — your namespace is empty because you haven’t deployed anything yet. You’re connected.

Next: Deploy an app.