OpenShift OCP 4.x — Installation Guide
Option 1: Red Hat Developer Sandbox (Free, No Install)
The fastest path to a real OpenShift environment with zero local setup, matching Roadmap's own recommendation:
1.Go to developers.redhat.com/developer-sandbox and sign up for a free account
2.A time-limited (renewable), real OpenShift namespace is provisioned automatically
3.Access via the web console or by installing the oc CLI locally and logging in with the provided token
bash
oc login --token=<sandbox-token> --server=<sandbox-api-url>
oc whoami
Option 2: OpenShift Local (`crc`) — Local Single-Node Cluster
For offline practice or deeper hands-on work with cluster-admin-level access:
bash
# Download crc from Red Hat's official site (requires a free Red Hat account
# for the pull secret)
crc setup
crc start
# This downloads and boots a single-node OpenShift cluster locally —
# expect this to take real time and meaningful local resources
crc console # Opens the web console
eval $(crc oc-env)
oc login -u developer https://api.crc.testing:6443
Install the `oc` CLI Standalone
If you're connecting to an existing cluster (Sandbox, a shared dev cluster) without needing crc:
bash
# Linux
curl -LO https://mirror.openshift.com/pub/openshift-v4/clients/ocp/latest/openshift-client-linux.tar.gz
tar xvzf openshift-client-linux.tar.gz -C /usr/local/bin oc
oc version
Full Cluster Installation (`openshift-install`) — Advanced/Production Only
Referenced in Fundamentals — this provisions a real multi-node production cluster and is genuinely heavyweight; not the recommended path for initial learning:
bash
# Requires an install-config.yaml, consumed ONCE during install
openshift-install create cluster --dir=./cluster-install
# This provisions Bootstrap, Control Plane, and Worker nodes per
# Overview's four-node-type architecture, then decommissions the
# Bootstrap node automatically once the Control Plane is self-sufficient
Verify Everything Works
bash
# 1. Confirm you're authenticated and can see cluster info
oc whoami
oc cluster-info
# 2. Confirm you can create and interact with a project (OCP's
# namespace-equivalent, with extra scaffolding)
oc new-project my-first-project
oc get pods -n my-first-project
# 3. Deploy a test application using oc's higher-level app-creation command
oc new-app --name=hello-openshift \
registry.access.redhat.com/openshift3/ose-hello-openshift
oc expose svc/hello-openshift # Creates a Route (not an Ingress)
oc get route hello-openshift
Common Installation Issues
crc start fails with insufficient resources
bash
crc config set memory 16384
crc config set cpus 6
# crc genuinely needs meaningful local resources — this is expected
# on an under-resourced machine, not a bug
oc login succeeds but every command fails with permission errors
bash
oc whoami
oc auth can-i create pods
# Confirm you're actually in the correct project/namespace scope —
# a common early mistake is assuming project-scoped RBAC works the
# same as cluster-admin access on a personal crc instance
Route created but not reachable externally
bash
oc get route hello-openshift -o yaml
# Confirm the Route's host actually resolves in your environment —
# on crc specifically, this typically requires crc's own DNS/hosts
# configuration to be set up correctly, not just the Route object existing
Developer Sandbox session expires
The free Sandbox is genuinely time-limited and needs periodic renewal (needs verification — recheck against current source for the current Sandbox duration/renewal policy) — this is expected behavior, not an error.
What's Installed
•oc CLI — the kubectl superset used for virtually all hands-on interaction in this guide
•A working OpenShift environment — either a free Sandbox namespace, a local crc single-node cluster, or (advanced) a full multi-node cluster via openshift-install
•Access to OCP-specific objects immediately available to practice with — Routes, Projects, and (with sufficient access) SCCs and Operators
Next Steps
Go to the Fundamentals section to explore OCP's architecture and create your first Route.
