OnCallReadyStart learning

OpenShift: glossary

100 terms, in plain words, with where each one is taught.

add-scc-to-user (oc adm policy add-scc-to-user)
oc adm policy add-scc-to-user SCC -z SA grants a ServiceAccount the right to use an SCC (a RoleBinding to the SCC's "use" ClusterRole). Lesson 47.9
anyuid
An SCC that allows any UID, root included. Priority 10, so it wins over restricted-v2 for every pod of a ServiceAccount that may use it. A last resort. Lesson 47.7
Application is not available
The router's own 503 page: it found no Route for the host and path, or the Route has no ready endpoints behind it.
apps wildcard (*.apps, apps domain)
The DNS name *.apps.<cluster> that points at the router, so every Route gets a working hostname without anyone touching DNS. Lesson 47.1
arbitrary UID (arbitrary UIDs, random UID)
OpenShift running your container as a UID the image has never heard of, with group 0. Images must work for any UID: files group-owned by 0 and group-writable. Lesson 47.7
ARO (Azure Red Hat OpenShift)
Azure Red Hat OpenShift: OpenShift run for you on Azure, managed jointly by Microsoft and Red Hat. Lesson 47.1
BuildConfig (bc)
The recipe for building an image inside the cluster: where the source is, which strategy, which builder image, and where to push the result. Lesson 47.1
builder image (builder images)
An image with a language runtime plus S2I scripts (assemble, run) that turn source code into a runnable app image, like nodejs:22-ubi9. Lesson 47.18
CatalogSource (catalog source, redhat-operators, certified-operators)
An index of operator packages OLM can install from, like redhat-operators. Lives in openshift-marketplace. Lesson 47.1
chroot /host
"Change root": inside oc debug node, makes /host the / of your shell so the node's own tools (crictl, journalctl, systemctl) run. Lesson 47.26
cluster operator (ClusterOperator, clusteroperators, oc get co)
An operator that manages one piece of the platform (DNS, ingress, the console, authentication...). Each reports Available / Progressing / Degraded. Lesson 47.1
Cluster Version Operator (CVO)
The operator that owns the cluster's version: during an upgrade it rolls every platform component forward to the new release. Lesson 47.1
ClusterServiceVersion (CSV (OLM), oc get csv)
ClusterServiceVersion: one installed version of an operator - its deployment, permissions and CRDs - with a phase (Succeeded, Failed). Lesson 47.1
ClusterVersion (oc get clusterversion)
The single object holding the cluster's version, update channel, available updates and upgrade history. Lesson 47.1
CR
Custom resource: one object of a CRD's kind, such as a Kafka named orders-events. An operator watches its CRs and does the work they ask for. Lesson 47.23
CRI-O
The container runtime OpenShift uses on its nodes (instead of containerd): it starts and stops containers for the kubelet. Lesson 47.1
DeploymentConfig (dc)
OpenShift's pre-Deployment workload object, with built-in triggers and hooks. Deprecated: migrate to Deployments. Lesson 47.1
Developer Sandbox
Red Hat's free OpenShift project on a shared cluster, for 30 days at a time, with no admin rights - a place to practise oc for real. Lesson 47.1
edge termination (edge)
A TLS Route where the router decrypts HTTPS and sends plain HTTP to the pod. The default for plain-HTTP apps. Lesson 47.1
EUS (Extended Update Support)
Extended Update Support: even-numbered OpenShift minor versions supported for longer, so you can upgrade less often. Lesson 47.1
HAProxy (HAProxy router)
A fast open-source load balancer and reverse proxy. OpenShift's router is HAProxy, configured from Route objects. Lesson 47.1
identity provider (IdP)
The system that actually knows users and passwords (Entra ID, LDAP, GitHub, an htpasswd file). The OAuth server asks it. Lesson 47.1
image change trigger (image trigger, image.openshift.io/triggers)
A rule that rolls a Deployment whenever an ImageStreamTag it uses moves to a new digest. Lesson 47.18
ImageStream (ImageStreams, is)
An OpenShift object that names images by tag and remembers which digest each tag points at - a pointer layer over registries. Lesson 47.1
ImageStreamTag (istag, istag/)
One tag of an ImageStream (web:latest) pointing at an exact image digest. Moving it to another digest is a deploy. Lesson 47.1
insecureEdgeTerminationPolicy (--insecure-policy, insecure policy)
What a TLS Route does with plain http://: None (503), Redirect (302 to https) or Allow (serve it). Lesson 47.13
install mode (installModes, OwnNamespace, AllNamespaces)
Which watch scopes an operator supports: OwnNamespace, SingleNamespace, MultiNamespace or AllNamespaces. The OperatorGroup must ask for a supported one. Lesson 47.23
InstallPlan (installplan)
The plan OLM makes for a Subscription: which CSV and CRDs to create. With Manual approval it waits until someone sets approved: true. Lesson 47.1
installPlanApproval (Manual approval, Automatic approval)
Automatic (OLM installs and upgrades by itself) or Manual (every InstallPlan waits for a human). Lesson 47.23
internal registry (image-registry.openshift-image-registry.svc)
The container image registry that runs inside every OpenShift cluster, where builds push their images. Lesson 47.1
kubeadmin
The temporary all-powerful user an OpenShift install creates. Use it once to set up a real identity provider, then delete its secret. Lesson 47.1
Kubernetes distribution (distribution of Kubernetes)
Kubernetes packaged by a vendor with chosen add-ons, defaults and support - like Ubuntu is a packaging of Linux. Lesson 47.1
Machine Config Operator (MCO, MachineConfig, MachineConfigPool)
The operator that owns every node's OS configuration and updates: it applies MachineConfigs and reboots nodes one pool at a time. Lesson 47.1
MCS label (MCS, s0:c26,c15)
The per-project SELinux category pair (like s0:c26,c15) OpenShift gives every pod, so a pod of one project cannot read another project's files even as the same UID. Lesson 47.4
minor upgrade (y-stream)
An upgrade to the next minor version (4.16 to 4.17): new features and removed APIs, so it needs preparation and an admin acknowledgement. Lesson 47.26
must-gather (oc adm must-gather)
oc adm must-gather collects logs, cluster operator state and resource definitions into a directory: the bundle Red Hat support asks for. Lesson 47.26
nonroot-v2
An SCC like restricted-v2 that lets the pod pick its own UID as long as it is not 0. Grant it for images that need a fixed non-root UID. Lesson 47.7
OAuth server (oauth-openshift)
The cluster's built-in login service: it checks your user against an identity provider and hands out the access tokens oc uses. Lesson 47.1
oc (oc CLI)
OpenShift's command-line tool. It contains all of kubectl plus OpenShift commands (login, new-project, new-app, expose, rsh, adm). Lesson 47.1
oc adm inspect
Collects everything about one namespace or object (YAML, events, logs) into a directory - a small, targeted must-gather. Lesson 47.26
oc adm upgrade
Shows the version, channel and recommended updates; --to=VERSION starts an upgrade; channel NAME switches channels. Lesson 47.26
oc debug node (oc debug node/)
Starts a privileged pod on a node with the node's filesystem at /host - the supported way to a node shell instead of SSH. Lesson 47.2
oc expose (oc expose svc)
oc expose svc/NAME creates a plain-HTTP Route for a Service (with a generated hostname unless --hostname). Lesson 47.2
oc login
Logs in to a cluster with a user and password (through the OAuth server) or --token=..., and writes the result into your kubeconfig. Lesson 47.1
oc new-app
Creates an app from an image, a Git repository (builder~repo) or a template: BuildConfig, ImageStream, Deployment and Service in one command. Lesson 47.2
oc policy add-role-to-user (add-role-to-user)
Gives a user a role in the current project: oc policy add-role-to-user edit alice creates the RoleBinding. Lesson 47.4
oc rsh
"Remote shell": oc rsh POD opens sh in the pod's first container - kubectl exec -it POD -- sh with less typing. Lesson 47.2
oc set triggers
Adds or removes the image change trigger on a Deployment: oc set triggers deploy/web --from-image=web:prod -c web. Lesson 47.21
oc start-build
Starts a new Build of a BuildConfig by hand (--follow to stream its log). Lesson 47.18
oc status (oc status --suggest)
Draws a project's Route -> Service -> Deployment -> BuildConfig chains and lists what is broken; --suggest explains each problem. Lesson 47.9
oc tag
Points an ImageStreamTag at another image: oc tag app:1.4 app:prod promotes (or rolls back) without rebuilding. Lesson 47.18
oc whoami (oc whoami -t, --show-server, --show-console)
Prints who you are logged in as; -t your token, --show-server the API URL, --show-context the kubeconfig context, --show-console the web console URL. Lesson 47.2
OCP
OpenShift Container Platform: the paid, supported OpenShift you install on your own machines or cloud. Lesson 47.1
OKD
The free community version of OpenShift, built from the same source as OCP. Lesson 47.1
OLM (Operator Lifecycle Manager)
Operator Lifecycle Manager: the part of OpenShift that installs, upgrades and wires up operators from catalogs. Lesson 47.1
OpenShift (OpenShift Container Platform)
Red Hat's Kubernetes distribution: Kubernetes plus a router, builds, an image registry, a web console, operators for everything and stricter security defaults, sold and supported as one product. Lesson 47.1
OpenShift Build (oc logs -f bc/)
One run of a BuildConfig, in a pod: its logs are oc logs -f bc/NAME, its output an image pushed to an ImageStreamTag. Lesson 47.18
OpenShift Dedicated
A managed OpenShift cluster run by Red Hat for one customer. Lesson 47.1
OpenShift project (Projects, oc projects, oc project)
A Kubernetes namespace with OpenShift metadata (display name, requester) and the rules for who may create one. oc project switches to one, oc projects lists yours. Lesson 47.1
OpenShift Route (route.openshift.io, Route object)
OpenShift's object for exposing a Service over HTTP(S) at a hostname through the router - the older, richer cousin of Ingress. Lesson 47.1
OpenShift router (the router, ingress controller (OpenShift), router-default)
The HAProxy pods in openshift-ingress that receive *.apps traffic and send it to pods, configured from every Route. Lesson 47.1
OpenShift Template (oc process, template.openshift.io)
The pre-Helm OpenShift package format: objects with ${PARAMETER} placeholders, rendered with oc process -p NAME=value. Lesson 47.29
openshift-operators
The namespace for operators that watch all namespaces; it already has an all-namespaces OperatorGroup. Lesson 47.23
openshift.io/scc
The annotation admission writes on every pod, naming the SCC that admitted it. oc get pod -o yaml | grep openshift.io/scc. Lesson 47.7
OperatorGroup (og)
The object in a namespace that says which namespaces the operators installed there watch. Exactly one per namespace. Lesson 47.1
OperatorHub
The catalog of operators an OpenShift cluster can install, shown in the console and backed by catalog sources. Lesson 47.1
PackageManifest (packagemanifest)
A read-only view of one operator package in a catalog: its channels, default channel, versions and install modes. Lesson 47.23
passthrough termination
A TLS Route where the router never decrypts: it forwards the encrypted stream by hostname (SNI) and the pod does TLS itself. Lesson 47.15
PHP
A programming language widely used for websites (WordPress is written in it). Lesson 47.18
privileged SCC (privileged (SCC))
The SCC that allows everything: root, host network, host paths, privileged containers. For node agents only. Lesson 47.7
project request template (project template)
The template a cluster admin sets that every new project is created from: quotas, limits, network policies, default role bindings. Lesson 47.4
ProjectRequest (oc new-project)
The object oc new-project creates: a request the cluster fulfils by making the namespace from a template and giving you admin on it - no cluster rights needed. Lesson 47.1
reencrypt termination (reencrypt)
A TLS Route where the router decrypts, applies HTTP features, and opens a new TLS connection to the pod. Lesson 47.1
ReplicationController (RC, RCs, ReplicationControllers)
The older Kubernetes object that keeps N copies of a pod running, before ReplicaSets. A DeploymentConfig creates one RC per revision. Lesson 47.21
resource category (oc get all)
A group name resource types sign up to. "all" is one: it covers pods, services, deployments, routes, builds... but not ConfigMaps, Secrets, PVCs or RoleBindings. Lesson 47.29
restricted-v2
The default SCC: a random UID from the project's range, no root, all capabilities dropped except NET_BIND_SERVICE, no host access, seccomp on. Lesson 47.1
RHCOS (Red Hat CoreOS, Red Hat Enterprise Linux CoreOS)
The operating system OpenShift nodes run: minimal, updated as one whole image by the cluster itself, not by hand with a package manager. Lesson 47.1
RHEL (Red Hat Enterprise Linux)
Red Hat Enterprise Linux: Red Hat's supported Linux distribution. RHCOS and the UBI images are built from it. Lesson 47.1
ROSA
Red Hat OpenShift Service on AWS: the managed OpenShift on Amazon's cloud. Lesson 47.1
route admission (Admitted)
The router accepting a Route: it shows Admitted in oc describe route. Two Routes claiming the same host and path: the older wins, the other is rejected. Lesson 47.4
route weights (alternateBackends)
A Route can split traffic between Services by weight (alternateBackends) - the router's built-in canary and blue-green. Lesson 47.13
rpm-ostree
The tool that updates RHCOS as one whole OS image instead of package by package, with the previous image kept for rollback. Lesson 47.26
S2I (Source-to-Image, source-to-image)
Source-to-Image: build an image from application source alone, using a builder image that knows the language - no Dockerfile. Lesson 47.1
SCC (SecurityContextConstraints, SCCs)
SecurityContextConstraints: OpenShift's admission rule sets for pods - which UIDs, capabilities, volumes and host access a pod may have. Pods are checked against the SCCs their ServiceAccount may use. Lesson 47.1
SCC priority
The order SCCs are tried in: higher priority first, then most restrictive first. The first SCC the pod fits admits it. Lesson 47.9
scc-subject-review (oc adm policy scc-subject-review)
oc adm policy scc-subject-review -z SA -f pod.yaml asks which SCC would admit the pod for that ServiceAccount, without creating it. Lesson 47.9
self-provisioner (self-provisioners)
The ClusterRole that lets users create their own projects. Removing it from all authenticated users makes project creation admin-only. Lesson 47.4
SELinux (SELinux label)
A Linux kernel security system that labels every process and file and allows only listed combinations. On by default on OpenShift nodes; each project gets its own label. Lesson 47.4
service CA (service-ca, serving-cert-secret-name)
OpenShift's internal certificate authority: annotate a Service and it writes a TLS certificate for the Service into a Secret, trusted inside the cluster. Lesson 47.15
sha256~ token (sha256~)
The format of an OpenShift OAuth access token. oc whoami -t prints yours; it expires (24 hours by default). Lesson 47.2
Subscription (OLM)
The object that says "install this operator package from this catalog, follow this channel, approve upgrades automatically or manually". Lesson 47.23
system:admin
The cluster-admin identity of the certificate-based kubeconfig an installer writes. Not an OAuth user: kept as break-glass access. Lesson 47.1
Thanos Querier (Thanos)
A front end that queries several Prometheus servers as one. OpenShift's console uses it to query platform and user metrics together. Lesson 47.29
UBI (Universal Base Image)
Red Hat's Universal Base Images: freely redistributable RHEL-based base images (ubi9, ubi9-minimal) built to run under arbitrary UIDs. Lesson 47.9
UID range (openshift.io/sa.scc.uid-range, sa.scc.uid-range)
The block of user ids (like 1000680000/10000) each project gets in an annotation. restricted-v2 runs pods with the first UID of that range. Lesson 47.1
update channel (channel)
A named stream of versions (stable, fast, stable-4.16) an operator or the cluster itself follows for upgrades. Lesson 47.23
usable SCC (use verb on securitycontextconstraints)
An SCC a ServiceAccount (or user) has RBAC permission to "use". Only usable SCCs are tried when the pod is admitted. Lesson 47.9
user workload monitoring (openshift-user-workload-monitoring)
An optional second Prometheus that scrapes ServiceMonitors in your own projects, alongside the platform's openshift-monitoring. Lesson 47.29
VIP (virtual IP)
Virtual IP: an address not tied to one machine; software like keepalived moves it to whichever node is healthy, so clients always reach a live one. Lesson 47.13
z-stream (z-stream upgrade, patch release)
An upgrade within one minor version (4.16.10 to 4.16.12): fixes only, the routine kind. Lesson 47.26