Skip to content
Hands-on IT IT Drills

Appendix A: Cheat Sheet

The most important commands from all days, sorted by exam domain. All of them were used live in the book; the days are given in parentheses.

Setup (Day 2)

source <(kubectl completion bash)
alias k=kubectl
complete -o default -F __start_kubectl k
export do="--dry-run=client -o yaml"
k config use-context <context>                       # first action of every task
k config set-context --current --namespace=<ns>
k explain deployment.spec.strategy --recursive
k api-resources --namespaced=false

Building and Operating the Cluster (Days 0, 1, 13, 16)

sudo kubeadm init --kubernetes-version=v1.37.2 --pod-network-cidr=192.168.0.0/16 \
  --apiserver-advertise-address=<IP>
kubeadm token create --print-join-command
sudo kubeadm init phase upload-certs --upload-certs   # key for control planes
kubeadm token create --print-join-command --certificate-key <KEY>
sudo kubeadm upgrade plan
sudo kubeadm upgrade apply v1.37.2 -y                        # first control plane
sudo kubeadm upgrade node                                    # all other nodes
k drain <node> --ignore-daemonsets --delete-emptydir-data
k uncordon <node>
sudo kubeadm certs check-expiration
sudo kubeadm certs renew all

Upgrade per node: package repository to the new minor version, apt-mark unhold, install kubeadm, kubeadm upgrade …, install kubelet and kubectl, apt-mark hold, systemctl daemon-reload, systemctl restart kubelet.

etcd (Days 13, 16)

sudo etcdctl --endpoints=https://127.0.0.1:2379 \
  --cacert=/etc/kubernetes/pki/etcd/ca.crt \
  --cert=/etc/kubernetes/pki/etcd/server.crt \
  --key=/etc/kubernetes/pki/etcd/server.key \
  snapshot save /var/lib/etcd-backup.db
sudo etcdutl snapshot status /var/lib/etcd-backup.db -w table
sudo etcdutl snapshot restore /var/lib/etcd-backup.db --data-dir /var/lib/etcd-restore
# then set the hostPath in the etcd manifest to /var/lib/etcd-restore
sudo etcdctl ... member list -w table
sudo etcdctl ... endpoint health -w table

RBAC (Day 3)

k create role <r> -n <ns> --verb=get,list,watch --resource=pods
k create rolebinding <rb> -n <ns> --role=<r> --user=<u>
k create rolebinding <rb> -n <ns> --clusterrole=view --serviceaccount=<ns>:<sa>
k create clusterrolebinding <crb> --clusterrole=<cr> --group=<g>
k auth can-i <verb> <resource> -n <ns> --as=system:serviceaccount:<ns>:<sa>
k certificate approve <csr>
k get csr <csr> -o jsonpath='{.status.certificate}' | base64 -d > user.crt
k create token <sa> -n <ns> --duration=10m

Several rules with different verbs → Role as YAML (Appendix B, Exercise 4).

Workloads and Scheduling (Days 4, 5, 6)

k create deployment <d> --image=<img> --replicas=3
k set image deployment/<d> <container>=<img>
k rollout status|history|undo deployment <d>
k scale deployment <d> --replicas=5
k create job <j> --image=busybox:1.37 -- date
k create job <j> --from=cronjob/<cj>
k create cronjob <cj> --image=busybox:1.37 --schedule='*/5 * * * *' -- date
k create configmap <cm> --from-literal=K=V
k create secret generic <s> --from-literal=K=V
k set resources deployment <d> --requests=cpu=100m --limits=memory=128Mi
k autoscale deployment <d> --cpu=60% --min=2 --max=6
k patch pod <p> --subresource resize -p '<json>'             # in-place resize
k taint node <n> key=value:NoSchedule        # with "-" at the end to remove it again
k label node <n> key=value                   # "key-" removes the label
k cordon|uncordon <n>
k top nodes --sort-by=cpu
k top pods -A --sort-by=memory

Static Pods: manifest into /etc/kubernetes/manifests (path: staticPodPath in /var/lib/kubelet/config.yaml). DaemonSet: generate Deployment YAML, change kind, delete replicas and strategy.

Services, Ingress, Gateway, DNS (Days 7, 8)

k expose deployment <d> --port=80 --target-port=8080 [--type=NodePort|LoadBalancer]
k create service nodeport <s> --tcp=80:8080 --node-port=30080
k set selector svc <s> app=<label>
k get endpointslices -l kubernetes.io/service-name=<s>
k create ingress <i> --class=traefik --rule="host/path*=svc:80,tls=<secret>"
k create secret tls <s> --cert=tls.crt --key=tls.key
k get gateway,httproute
k get cm coredns -n kube-system -o jsonpath='{.data.Corefile}'

DNS names: <svc>.<ns>.svc.cluster.local, Pods <ip-with-dashes>.<ns>.pod.cluster.local.

NetworkPolicies (Day 9)

from:                           from:
- namespaceSelector: {...}      - namespaceSelector: {...}
- podSelector: {...}              podSelector: {...}
  -> OR (two entries)             -> AND (one entry)

Default deny: podSelector: {} and only policyTypes. With egress, allow DNS (UDP and TCP 53). Ingress controllers need their own allow rule.

Storage (Day 10)

k get pv,pvc,sc
k patch pv <pv> -p '{"spec":{"claimRef":null}}'          # Released -> Available
k patch pvc <pvc> -p '{"spec":{"resources":{"requests":{"storage":"2Gi"}}}}'

PVC Pending: check class, access mode and size against the PVs. The storageClassName of a PVC is immutable.

Helm, Kustomize, Extensions (Days 11, 12)

helm repo add <name> <url> && helm repo update
helm install <rel> <chart> --version <v> -n <ns> --create-namespace \
  --set k=v -f values.yaml
helm upgrade <rel> <chart> --reuse-values --set k=v
helm history|rollback|uninstall <rel>
helm template <rel> <chart> | grep kind:
k kustomize <overlay>
k apply -k <overlay>
k get crd
k api-resources --api-group=<group>
sudo crictl ps -a | pods | images | logs <id> | inspect <id>

Troubleshooting (Days 14, 15)

Symptom First command
Node NotReady systemctl status kubelet, journalctl -u kubelet
container runtime is down systemctl status containerd
Pod Pending, no events check the scheduler Pod
Deployment without ReplicaSet check the controller manager Pod
connection refused to the API kubeconfig port, crictl ps -a, crictl logs
CrashLoopBackOff k logs, k logs --previous
OOMKilled, exit 137 memory limit
Service refused k get endpointslices
bad address CoreDNS
timed out NetworkPolicy
k get events --sort-by=.lastTimestamp
k get --raw='/readyz?verbose'
k logs <pod> -c <container> --tail=20 --since=5m
k logs -l app=<label> --prefix --tail=1
sudo ls /var/log/pods/