The machine was the easy part.
Now watch how software reaches it.

A note on the commands: this site uses oc, OpenShift's CLI. The Kubernetes commands here work the same under kubectl; oc is a superset, and the OpenShift-only parts (SCCs, Routes) are its own. Read kubectl in the docs, type whichever your cluster gives you.

That 1/1 reads as containers-ready over containers-wanted: a pod can hold more than one, which is Course IIIb.

An image got built - of what, exactly? Pushed - to where, and what travelled? Applied - which is not the same as launched. Running - according to whom? Every course below takes one of those words apart. The armour is already loose.

A container is a process, not a machine

If you arrived here from Linux, take this translation first: a running container is an ordinary process on your kernel. No guest OS, no hypervisor. The kernel gives it namespaces so it sees its own PID tree, mounts, network and hostname, and cgroups so its CPU and memory can be capped. ps on the host lists it. kill on the host kills it. Podman leans into that: no daemon sits in the middle - the container is a child of your own shell - and rootless mode maps your user onto root inside the container through a user namespace, so root in there is an unprivileged UID out here.

So what does the image provide? The filesystem that process sees. That is the whole job, and it is why an image is a stack of layers rather than a disk image.

Field note. Prove it on your own box. podman run -d --name t alpine sleep 300, then on the host: ps -ef | grep sleep finds the process, lsns -p <pid> lists the namespaces it was handed, cat /proc/<pid>/cgroup shows where its limits live, and mount | grep overlay shows the layers stitched together. None of it is exotic. It is your kernel, described differently.

What an image is made of

An image is not a copy of a machine. It is a stack of read-only layers, each recording only what changed from the one beneath. Layers are content-addressed, so an identical layer is stored once and reused by every image that references it, so a pull only fetches the layers you do not already have. A base sits at the bottom, your dependencies on it, your code - usually the smallest layer, always the most volatile - above that.

Those layers become one filesystem through a union mount - overlayfs, the same kernel feature you can mount by hand. The read-only image layers are the lower dirs; the container gets a fresh upper dir of its own. Writes land in the upper, and editing an existing file copies it up there first, leaving the image layer untouched underneath.

That upper dir is the top slab in the scene, and it is the odd one out: the writable layer is not part of the image. The runtime creates it with the container and discards it when that container dies. Nothing written there ships, and nothing written there survives - which is the entire reason volumes exist.

Change a layer and every layer above it must be rebuilt.

The order is a caching decision

The builder caches layer by layer, and a cached layer survives only while everything beneath it is unchanged. Put COPY . . above your dependency install and you have told the builder to throw the dependency cache away every time one line of code changes. Dependencies first, code last - a Containerfile is a cache policy that happens to build software.

Field note. A rebuild that takes twenty minutes and one that takes twenty seconds are usually the same Containerfile with the lines swapped.

This stack explains the whole ecosystem above it: sharing explains why pulls are fast, immutability explains why a digest can name the exact bytes (Course VII), and the throwaway top layer explains why state needs volumes. One idea - frozen diffs - all the way down.

Pre-reads: none - start here  Further: Podman get-started · Podman docs + builds

Desired against actual, on a loop

The habit underneath everything: the reconciliation loop - compare desired state against actual state, fix the difference, repeat. That oc apply didn't launch anything; it filed paperwork. The machine took it from there, and it never stops taking it from there: kill a pod and it returns, not because something noticed the crash but because the loop noticed the difference.

Kubernetes doesn't run your app - it reconciles it.

One half decides, one half runs

The split in the scene is the split that makes everything else possible: a control plane that decides - holds the truth, schedules, reconciles - and worker nodes that run pods. The workers are deliberately interchangeable: identical, replaceable, cattle from day one. Authority does not live where the work happens.

Field note. If you are SSHing into nodes to "fix" things, you are arm-wrestling the reconciler - and it does not get tired. Change the desired state instead.

What the control plane is made of

"Control plane" is four processes and a database, and naming them makes every later error message readable. The api-server is the only door: everything authenticates, is authorised and is admitted there, and it is the only component allowed to touch etcd - the key-value store holding the entire cluster state. Lose etcd and you have lost the cluster, which is why backing it up is the homework nobody should skip. The scheduler decides which node a new pod belongs on - packing by the resources a pod requests, not by what it currently uses - and writes that decision down; it never starts anything. The controller-manager runs the reconciliation loops. Nothing talks sideways - every component watches the api-server.

One of those loops is the chain you will debug most: a Deployment creates a ReplicaSet, and the ReplicaSet creates pods. That is why the deploy in the aperitif printed deployment.apps/api created and you then went looking for a pod - and why, when a rollout is stuck with no pod at all, the answer is upstream in that chain rather than on any node.

etcd is the truth. Everything else is a cache of it.

Everything the delivery arc teaches from Course VI onward is the same loop at a bigger scale - git as the desired state, whole fleets as reconciled objects. Learn the promise once; it repeats all the way up.

Pre-reads: C-I  Further: Kubernetes overview · cluster architecture

Where a pod becomes processes

Every node runs a kubelet - the agent that owns what should be running there. It does not create containers itself. It speaks CRI - the Container Runtime Interface - to containerd or CRI-O, and that runtime pulls the image (through the mirror of Course VII) and actually creates and starts the container, handing the low-level work to runc or crun. A CNI plugin (Container Network Interface) hands the pod a real IP, and kube-proxy - or an eBPF datapath (code running inside the kernel itself) replacing it - makes Service addresses route to real pods. One correction worth carrying: the runtime calls the CNI plugin, not the kubelet.

The kubelet decides what should run. The runtime is what starts it.

Field note. Node NotReady? The kubelet is a systemd unit like any other: journalctl -u kubelet on that node, and check it can still reach the api-server. A node that cannot phone home is presumed lost.

The control plane never touches your workload. It writes intent; the kubelet turns intent into instructions, and the runtime turns instructions into processes. Authority and execution meet exactly here, nowhere else.

Pre-reads: C-II  Further: cluster architecture · node components

The jacket, not the container

Everything inside the jacket shares a network namespace: one IP, localhost between friends. Volumes are declared once on the pod, but each container mounts the ones it needs - sharing storage is opt-in, not automatic. initContainers run first, in order, each to completion - nothing else in the pod starts until every one of them has exited successfully. Sidecars ride along with their own containers and their own jobs: proxy, logs, reload.

Three probes, three different jobs

startup owns warm-up, readiness gates traffic, liveness restarts the truly hung. Confusing them is how healthy pods get executed - a slow start killed by an impatient liveness probe looks exactly like a crash.

containerPort is documentation - unless a Service targets it by name, or you use hostPort. Either way the app still has to bind the port itself.

Field note. Exit 137 is 128 + 9: the process was SIGKILLed. It does not say by whom. The kernel's OOM killer surfaces as reason OOMKilled; a failed liveness probe shows up in the pod's events; eviction and node pressure look different again. One exit code, several possible crimes - read the termination reason and the events, never the number alone: oc describe pod <name> shows both together, and oc logs --previous shows what the dead container said on its way out. Worth connecting to what you already know: a memory limit becomes a cgroup ceiling, and the kernel's OOM killer enforces it exactly as it would for any other process on the box.

The pod is the smallest schedulable unit - the jacket, never the container. Once that distinction lands, half of Kubernetes networking stops being mysterious.

Pre-reads: C-III's node scene above  Further: pods · the three probes

The stable name in front of the churn

A Service is the fixed point: a ClusterIP inside, a LoadBalancer at the edge, Ingress or the Gateway API doing host- and path-routing above. Clients hold the name; the pods behind it come and go without anyone being told.

How a Service finds its pods

A Service holds no list of pods. It holds a label selector - match app: api - and a controller continuously matches that against every pod in the namespace - a Kubernetes namespace this time, a naming boundary for objects, no relation to the kernel namespaces of Course I - keeping the passing set in an EndpointSlice. Labels are how everything here finds everything else, from a Service picking pods to a fleet hub picking whole clusters (Course VI). Wear the label and you are eligible; readiness decides whether you stay.

And the ClusterIP is worth a Linux translation: no interface owns that address. Nothing answers ARP for it. It exists as a rule - kube-proxy programs iptables (or IPVS) on every node so packets aimed at the virtual IP are DNATed to one of the ready pod IPs, and an eBPF dataplane does the same job further down without the rule tables. Cluster DNS - CoreDNS - resolves api.myns.svc.cluster.local to that VIP. If you have ever written a DNAT rule by hand, you have already built a small Service.

Readiness decides membership

A pod failing its readiness probe silently leaves the pool. No error, no event at the client - traffic simply stops arriving. That is the feature: broken instances remove themselves. It is also the first place to look when traffic "disappears".

The rule has deliberate exceptions: you address pods directly when debugging a specific instance, and headless Services exist precisely so StatefulSet members can be reached individually by stable DNS name. For ordinary application traffic, though, the Service is the only address worth knowing.

For application traffic: never talk to a pod; talk to a Service.

Field note. "The network is broken" after a deploy is usually readiness telling the truth about your app - not the network lying about your packets.

The dark pod in the scene is not an error state - it is the system working. Membership is re-earned every few seconds, by every pod, for as long as it serves.

Pre-reads: C-III's pod scene (readiness lives there)  Further: Services and networking

The doorman interviews every pod

The SCC - Security Context Constraint - is admission deciding what a pod may BE, checked by the api-server when the pod is created and before any node sees it. The default, restricted-v2, runs your container as a random non-root UID - your image has to cope. Workloads that genuinely need privilege get a dedicated ServiceAccount bound to a minimal custom SCC, never the stock one.

The platform runs itself

Routes predate Ingress and still rule here. An Operator is a controller paired with a custom resource: you describe what you want in YAML, and its controller builds it and keeps it true - the reconciliation loop of C-II, sold as a product. The split matters: the platform's own operators are driven by the Cluster Version Operator, while OLM installs and upgrades the add-on Operators you choose from OperatorHub. The OS underneath is immutable - changed by MachineConfig, never by SSH. And SNO - single-node OpenShift - puts the whole cluster on one box at the edge. At fleet scale the labels from Course VI decide which of these boxes runs what.

On OpenShift, admission is the interview - the SCC is the dress code.

Field note. Deployment stuck at 0/1 with no pod at all? The refusal happened above scheduling - read the ReplicaSet events. The error lives a level up.

Everything in the ring wraps the same core you already know. What OpenShift adds is a house style with teeth - and admission is where it bites first.

Pre-reads: C-II  Further: OpenShift documentation · Red Hat OpenShift

The loop you already know, one level up

C-II taught the reconciliation loop: desired versus actual, fix the difference, repeat. ArgoCD is the same habit applied to delivery. An Application names a repo, a path and a revision - watch this branch of this repository - and the controller renders what it finds there, compares it against the live cluster, and syncs the difference. The deploy button is a git commit; the change history is git log; code review is change control. Git is the record of intent - the cluster's own audit log and the reconciler's sync history still record what actually happened, including everything git never sees.

oc is for archaeology. git is for change.

Pull, not push - the security inversion

Here the cluster pulls. Push-based delivery exists and is still GitOps to many - this is the stronger variant, and worth choosing deliberately: no CI system, no laptop, no build pipeline holds a credential that can touch the cluster, because the agent inside holds a read-only deploy key and the trust arrow points out. Compromise the build system and you can propose a change, which is visible; you cannot reach into production through git. It can still push images - which is the other half of why a manifest should name the digest, not the tag. Hand-edit a live object and the controller flags it OutOfSync - with selfHeal enabled it puts the object back, and with prune enabled what leaves git leaves the cluster. Both are opt-in: without them the reconciler reports the drift and waits for a human. Rollback is git revert, which is why commit hygiene is an operational skill.

Field note. A hand-patch on a live object survives exactly until the next sync. If you must touch production directly to stop the bleeding, open the pull request in the same hour - otherwise the reconciler quietly undoes your fix, and the outage returns with nobody able to say why.

At fleet scale, the label is the deploy button

Run many OpenShift clusters under a hub - RHACM, Red Hat's fleet manager, with edge clusters arriving through zero-touch provisioning - and nobody applies apps to clusters by hand. Each app carries a Placement that selects cluster labels; the hub matches placements against the labels a cluster wears, and the chosen cluster is given its assignment - push-style from the hub by default, or in a pull model where each cluster runs its own reconciler, which is the variant this arc teaches. Either way, labelling the cluster is the deploy action: attach the label and the app follows, remove it and the app leaves. The same pull model as above, one level bigger - a cluster's labels are its entitlements, reconciled like everything else.

Label the cluster; the app follows.

The one thing git never holds

Git holds everything except secrets. Commit a plaintext secret and you should treat it as compromised from that moment: deleting it later does not guarantee it is gone from history, forks, clones, CI caches or backups. So the pattern splits the reference from the value: git carries an ExternalSecret naming a logical key; a vault - Azure Key Vault in the worked example - holds the value; an operator inside the cluster exchanges one for the other at runtime. Rotation happens in the vault, never as a commit. That is why the vault floats apart in the scene above: it is never absorbed into the pipeline.

Git holds the shape of the secret. The vault holds the secret.

Delivery stops being an event and becomes a property: the cluster is always converging on what the repository says. "Who deployed this?" becomes "who merged this?" - and that question always has an answer.

Pre-reads: C-II · Kubernetes concepts · git + pull requests  Further: Argo CD · OpenShift GitOps · External Secrets · Azure Key Vault

Say the name properly

Three ways to name an image, in rising order of honesty: :latest (a moving target), :1.4.2 (a label someone maintains, until they re-push it), and name:1.4.2@sha256:... - a fact. The tag stays for human eyes; the digest does the pulling. Pin by digest and "what is running?" has exactly one answer.

Field note. :latest is how two nodes run different code from one manifest - the second node pulled an hour later, after a re-push. Nobody changed the YAML.

Why a fleet pulls once

Between the build and the node sits the registry chain. Upstream, a managed registry - Azure Container Registry in the worked example - holds what CI built. In front of the cluster sits a mirror: a pull-through cache like zot. The fleet asks the mirror, the mirror asks upstream once, everything after is local. Rate limits, egress cost, disconnected sites, control - one place to gate and audit what enters. OpenShift formalises the re-route with image mirror rules, and carries a sharp edge: digest-mirror rules rewrite digest pulls only, so a by-tag pull silently skips them - unless you also add an ImageTagMirrorSet, which is the rule type built for tag pulls. Pinning by digest is still the habit that makes the digest rules catch everything.

At real fleet scale the mirror itself tiers: a central mirror in the cloud fronts upstream once, and every site's mirror pulls from the centre rather than from upstream directly. A new image ripples outward in layers - upstream to the centre, centre to each site as it asks, site to its nodes over the LAN - instead of every site hammering upstream at the same moment. Upstream sees one consumer; each site sees one hop; the nodes never leave the building.

Mirrors tier: the load fans out in layers, never all at once.

Build once, promote by copy

Every rebuild is a new artefact - in practice a different digest (reproducible builds are the deliberate exception), untested by the stages before it. So build once, then promote the same digest through environments by copying, registry to registry - dev proves the exact bytes prod will run. On a multi-arch image that copy (skopeo copy) needs --all (and --preserve-digests to fail loudly rather than quietly), or you copy one architecture and the digest you promoted is not the digest that lands. Human tags ride along; the digest is the through-line.

If the digest changed, it is not a promotion - it is a new candidate.

Names that can move are convenient exactly until they move. Address content by what it is, and the supply chain stops resting on trust: it rests on a hash anyone can check.

Pre-reads: C-I · Kubernetes images  Further: zot · Azure Container Registry · OpenShift image mirroring · skopeo

You are writing a program, not YAML

Helm templates are Go text/template: {{ .Values.device.address }} is a pipeline walking a values object, _helpers.tpl holds the named templates - the partials every manifest includes, not functions you can call bare. You are not writing YAML - you are writing a program whose output is YAML. So render locally, read the output, and lint what came out, not what went in.

Review the render, not just the template.

Field note. ArgoCD renders charts with helm template rather than running helm install, so the lifecycle differs from Helm's own. lookup comes back empty - there is no live cluster at render time. Hooks are not dead, though: Argo maps Helm hooks onto its sync phases (pre-install and pre-upgrade become PreSync, post-install and post-upgrade become PostSync), while a few - rollback and test hooks - have no equivalent at all. Render the way your deployer renders, and check where your hooks actually land.

Contexts: the cluster's whole voice is one small file

The chart owns everything structural - resources, probes, security, policy. Each cluster owns one values file: names, addresses, sizes, flags. The context is deliberately values-only; the moment it carries its own manifests there are two owners for one object, and they will disagree. One value can feed many rendered artefacts - an address appearing in the app config, the network attachment and two policies renders from one field, so the copies cannot diverge - the lived version is on the blog.

The chart owns the shape. The context owns the numbers.

Make the template refuse

A template that renders whatever it is given just moves the failure downstream. The grown-up chart carries a values.schema.json: a context missing a required value fails at render time, in the pipeline, with a message naming the field - not months later as enforcement pointed at nothing.

Field note. The failure you want is the render that refuses. It costs a red pipeline. The alternative reports healthy the whole time.

Consistency stops being something you police and becomes something the tooling cannot express.

Pre-reads: C-II · Kubernetes objects  Further: Helm docs · chart template guide · Go text/template · Helm on OpenShift