Crear y escalar

Create and scale

Creare e scalare

# Crear un deployment# Create a deployment# Creare un deployment
k create deploy web --image=nginx:1.25 --replicas=3

# Escalar# Scale# Scalare
k scale deploy web --replicas=5

# Ver el deployment, sus RS y pods# See the deployment, its RS and pods# Vedere il deployment, i suoi RS e i pod
k get deploy,rs,pods

Rollout: actualizar la imagen

Rollout: update the image

Rollout: aggiornare l'immagine

# Cambiar la imagen: <container>=<imagen># Change the image: <container>=<image># Cambiare l'immagine: <container>=<immagine>
k set image deploy/web nginx=nginx:1.26

# Seguir el rollout (BLOQUEA hasta terminar o fallar)# Follow the rollout (BLOCKS until done or failed)# Seguire il rollout (BLOCCA fino a fine o errore)
k rollout status deploy/web

# Dejar registrado el motivo (--record está deprecado)# Record the reason (--record is deprecated)# Registrare il motivo (--record è deprecato)
k annotate deploy/web kubernetes.io/change-cause="update a 1.26"
k rollout history deploy/web

Rollback: volver atrás

Rollback: go back

Rollback: tornare indietro

k rollout undo deploy/web
k rollout undo deploy/web --to-revision=2

DaemonSet — uno por nodo

DaemonSet — one per node

DaemonSet — uno per nodo

# NO existe 'kubectl create daemonset'.# There is NO 'kubectl create daemonset'.# NON esiste 'kubectl create daemonset'.
# Truco: generá un deployment y editá el kind# Trick: generate a deployment and edit the kind# Trucco: genera un deployment e cambia il kind
k create deploy logger --image=fluentd $do > ds.yaml
#   kind: Deployment → DaemonSet · borrar replicas y strategy#   kind: Deployment → DaemonSet · drop replicas and strategy#   kind: Deployment → DaemonSet · togli replicas e strategy

k get daemonsets -A
# calico-node y kube-proxy ya son DaemonSets# calico-node and kube-proxy already are DaemonSets# calico-node e kube-proxy sono già DaemonSet

StatefulSet — identidad estable

StatefulSet — stable identity

StatefulSet — identità stabile

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: web-sts
spec:
  serviceName: nginx-h       # ↔ headless service (clusterIP: None)# ↔ headless service (clusterIP: None)# ↔ headless service (clusterIP: None)
  replicas: 3
  selector:
    matchLabels: { app: web-sts }    # ↔ labels del template# ↔ template labels# ↔ label del template
  template:
    metadata:
      labels: { app: web-sts }
    spec:
      containers:
      - name: nginx
        image: nginx
# pods: web-sts-0, -1, -2 (ordinales, en orden, sin hash)# pods: web-sts-0, -1, -2 (ordinal, in order, no hash)# pod: web-sts-0, -1, -2 (ordinali, in ordine, senza hash)

Crear

Create

Creare

# ConfigMap desde literales# ConfigMap from literals# ConfigMap da literal
k create configmap app-cfg \
  --from-literal=COLOR=blue \
  --from-literal=MODE=prod

# ConfigMap desde un archivo# ConfigMap from a file# ConfigMap da un file
k create configmap app-cfg --from-file=config.txt

# Secret (base64, NO cifrado por defecto)# Secret (base64, NOT encrypted by default)# Secret (base64, NON cifrato di default)
k create secret generic db-pass \
  --from-literal=password=s3cr3t

Manifiesto completo · las 4 formas juntas

Complete manifest · all 4 forms at once

Manifest completo · le 4 forme insieme

apiVersion: v1
kind: Pod
metadata:
  name: demo
spec:
  containers:
  - name: app
    image: nginx
    # ─── (A) CM/Secret COMPLETOS como env ───# ─── (A) WHOLE CM/Secret as env vars ───# ─── (A) CM/Secret INTERI come env ───
    envFrom:
    - configMapRef:
        name: app-cfg          # cada key del CM → env var (mismo nombre)# every CM key → env var (same name)# ogni key del CM → env var (stesso nome)
    - secretRef:
        name: db-pass          # cada key del Secret → env var# every Secret key → env var# ogni key del Secret → env var
    # ─── (B) UNA key puntual como env (renombrando) ───# ─── (B) ONE specific key as env (renamed) ───# ─── (B) UNA key precisa come env (rinominata) ───
    env:
    - name: APP_COLOR           # nombre que VOS elegís# the name YOU pick# il nome che scegli TU
      valueFrom:
        configMapKeyRef:
          name: app-cfg        # ← ConfigMap app-cfg# ← ConfigMap app-cfg# ← ConfigMap app-cfg
          key: COLOR            # ← su key COLOR# ← its COLOR key# ← la sua key COLOR
    - name: DB_PASSWORD
      valueFrom:
        secretKeyRef:
          name: db-pass        # ← Secret db-pass# ← Secret db-pass# ← Secret db-pass
          key: password         # ← su key password# ← its password key# ← la sua key password
    # ─── (C) montar como ARCHIVOS ───# ─── (C) mount as FILES ───# ─── (C) montare come FILE ───
    volumeMounts:               # DENTRO del container# INSIDE the container# DENTRO il container
    - name: cfg-vol           # ↔ enlaza con volumes[].name de abajo# ↔ links to volumes[].name below# ↔ si lega a volumes[].name qui sotto
      mountPath: /etc/config   # keys de app-cfg = archivos acá# app-cfg keys = files in here# le key di app-cfg = file qui
    - name: sec-vol
      mountPath: /etc/secret
      readOnly: true
  volumes:                     # nivel POD (hermano de containers)# POD level (sibling of containers)# livello POD (fratello di containers)
  - name: cfg-vol             # ↔ el volumeMount cfg-vol# ↔ the cfg-vol volumeMount# ↔ il volumeMount cfg-vol
    configMap:
      name: app-cfg            # ConfigMap: se usa  name:# ConfigMap: uses  name:# ConfigMap: usa  name:
  - name: sec-vol
    secret:
      secretName: db-pass      # Secret: OJO → es  secretName:# Secret: CAREFUL → it is  secretName:# Secret: ATTENZIONE → è  secretName:

Crear un autoscaler

Create an autoscaler

Creare un autoscaler

# El deployment DEBE tener requests.cpu primero# The deployment MUST have requests.cpu first# Il deployment DEVE avere prima requests.cpu
k set resources deploy web --requests=cpu=100m

# Escalar entre 2 y 10 réplicas, objetivo 50% de CPU# Scale between 2 and 10 replicas, target 50% CPU# Scalare fra 2 e 10 repliche, obiettivo 50% di CPU
k autoscale deploy web --min=2 --max=10 --cpu=50%
#   --cpu-percent=50 sigue andando pero está deprecado#   --cpu-percent=50 still works but is deprecated#   --cpu-percent=50 funziona ancora ma è deprecato

k get hpa
k describe hpa web

El <unknown> y cómo se arregla

The <unknown> and how to fix it

L'<unknown> e come si risolve

k get hpa            # TARGETS <unknown>/50%# TARGETS <unknown>/50%# TARGETS <unknown>/50%
k top pods           # el test de 3 segundos: ¿hay métricas?# the 3-second test: are there metrics?# il test di 3 secondi: ci sono metriche?

# Si no las hay, falta metrics-server (en labs kubeadm# If not, metrics-server is missing (in kubeadm labs# Se non ci sono, manca metrics-server (nei lab kubeadm
# además hay que tolerar los certs self-signed del kubelet)# you must also tolerate the kubelet's self-signed certs)# bisogna anche tollerare i cert self-signed del kubelet)
k patch deploy metrics-server -n kube-system --type=json \
  -p='[{"op":"add","path":"/spec/template/spec/containers/0/args/-","value":"--kubelet-insecure-tls"}]'

Para tu chuleta

For your cheat sheet

Per il tuo bigliettino

### Deployments
k create deploy NAME --image=IMG --replicas=N
k scale deploy NAME --replicas=N
k set image deploy/NAME cont=img:tag
k rollout status|history|undo deploy/NAME
k rollout restart deploy/NAME

### ConfigMaps / Secrets
k create configmap NAME --from-literal=K=V
k create secret generic NAME --from-literal=K=V
k exec POD -- printenv K

### HPA
k autoscale deploy NAME --min=2 --max=10 --cpu=50%

### Sin generador imperativo → dry-run y editar kind### No imperative generator → dry-run and edit the kind### Senza generatore imperativo → dry-run e cambia kind
#   DaemonSet · StatefulSet
k create service clusterip NAME --clusterip=None   # headless# headless# headless