Finding your way around a cluster

When you first land on an OpenShift cluster — yours or someone else’s — you need four things fast: the console URL, the API endpoint, the admin credentials, and the cluster’s real name. Here’s how to get each without clicking around.

Watch on YouTube ↗

Console and API URLs

# Web console
oc whoami --show-console
# https://console-openshift-console.apps.my-cluster.example.com

# API server
oc whoami --show-server
# https://api.my-cluster.example.com:6443

--show-console is the one people forget exists and instead go hunting through the web UI for. It reads the console route directly.

Who am I, and what can I do?

oc whoami                    # current username
oc whoami -t                 # current session token
oc auth can-i create pods    # does my account have a given permission?

The cluster’s real name

The name you typed at install time isn’t quite the cluster’s canonical identity — OpenShift derives an infrastructure name (with a random suffix) that tags cloud resources and appears in machine names:

oc get infrastructure cluster -o jsonpath='{.status.infrastructureName}'
# my-cluster-x7k2p

Tip

That suffix is why two clusters both called my-cluster never collide on cloud resources — the infrastructure name is unique per install.

The admin (kubeadmin) password

Where this lives depends on how the cluster was created:

Installer

The installer writes it once to the install directory:

cat <install-dir>/auth/kubeadmin-password

OpenShift Local (CRC)

crc console --credentials
cat ~/.crc/machines/crc/kubeadmin-password

Caution

kubeadmin is a temporary break-glass account. On any real cluster you should wire up a proper identity provider (OIDC, LDAP, htpasswd for labs) and remove kubeadmin once other cluster-admins exist. Don’t build workflows around it.

Cluster version and health at a glance

oc get clusterversion            # version + whether an update is progressing
oc get co                        # cluster operators — all should be Available
oc get nodes                     # every node should be Ready

If oc get co shows anything Degraded or Progressing for a long time, that’s usually where a cluster problem is hiding — start there.