Kubernetes

K3s Cluster Forming (Server & Agent Nodes)

K3s makes it easy to form a lightweight Kubernetes cluster by running a server node (control plane) and connecting agent nodes (workers) to it.

Prerequisite:

All nodes must have network connectivity to each other.
Install script: https://get.k3s.io
Required ports (6443, 8472/UDP, etc.) must be open.

Command Syntax for Server (Control Plane):

curl  -sfL  https://get.k3s.io | sh  -s  -  server  \

--node-name <server-name> \

--cluster-init

Parameters & Options (Server):

Parameter	Type	Description
`--node-name`	flag	Custom name for the server node
`--cluster-init`	flag	Initializes a new K3s cluster (1st node)

Common Use Case:

curl  -sfL  https://get.k3s.io | sh  -s  -  server  --node-name  master-1  --cluster-init

Sample Output:

[INFO] Starting k3s server...

[INFO] k3s is up and running on https://<server-ip>:6443

Command Syntax for Agent (Worker Node):

curl  -sfL  https://get.k3s.io | K3S_URL=https://<server-ip>:6443  \

K3S_TOKEN=<token> \

sh  -s  -  agent  \

--node-name <agent-name>

Parameters & Options (Agent):

Parameter	Type	Description
`K3S_URL`	env	API server URL (control plane)
`K3S_TOKEN`	env	Cluster join token from server
`--node-name`	flag	Name of the agent node

Get the Token (on Server):

sudo  cat  /var/lib/rancher/k3s/server/node-token

Example (Agent):

curl  -sfL  https://get.k3s.io | K3S_URL=https://192.168.1.10:6443  \

K3S_TOKEN=K1057c7... \

sh  -s  -  agent  --node-name  worker-1

Sample Output (Agent):

[INFO] Starting k3s agent...

[INFO] Connecting to https://192.168.1.10:6443

[INFO] k3s agent is up and running

Managing K3s Service with systemctl

Use these commands to check the status, restart, stop, or start the K3s service on your node.

Helpful for troubleshooting, service maintenance, or applying configuration changes.

Prerequisite:

You need sudo privileges on the node where K3s is installed.

Command Syntax & Usage:

Command	Description
`sudo systemctl status k3s`	Check the current status of K3s service
`sudo systemctl restart k3s`	Restart the K3s service
`sudo systemctl stop k3s`	Stop the K3s service
`sudo systemctl start k3s`	Start the K3s service

Common Patterns or Use Cases:

Check if K3s is running:

sudo  systemctl  status  k3s

Restart K3s to apply config changes or recover from issues:

sudo  systemctl  restart  k3s

Stop K3s service when maintenance is needed:

sudo  systemctl  stop  k3s

Start K3s after it has been stopped:

sudo  systemctl  start  k3s

Sample Output for status:

● k3s.service - Lightweight Kubernetes

Loaded: loaded (/etc/systemd/system/k3s.service; enabled; vendor preset: enabled)

Active: active (running) since Tue 2025-05-28 09:10:01 UTC; 2h 15min ago

Main PID: 1234 (k3s-server)

Tasks: 34 (limit: 4915)

Memory: 250.0M

CPU: 3min 22.567s

Applying Kubernetes Configuration in K3s

Applies or updates resources defined in a YAML/JSON manifest file.

Used to create, update, or delete Kubernetes resources declaratively.

Prerequisite:

Have the manifest file (file.yaml) ready locally or accessible by your terminal.

Ensure correct namespace is specified if resource is namespace-scoped.

Command Syntax:

kubectl  apply  -f <file.yaml> -n <namespace>

Parameters & Options:

Parameter	Type	Description
`-f <file.yaml>`	flag	Path to the YAML or JSON manifest file
`-n`	flag	Namespace to apply the resource to

Common Patterns or Use Cases:

Apply a deployment manifest to a namespace:

kubectl  apply  -f  deployment.yaml  -n  app

Sample Output:

deployment.apps/myapp created

service/myservice configured

configmap/myconfig unchanged

Getting Namespaces in K3s

List all namespaces in the cluster. Namespaces help organize and isolate resources.

Command Syntax:

kubectl get ns

Parameters & Options:

Parameter	Type	Description
`ns`	arg	Short form for `namespace`

Common Use Case:

List all namespaces:

kubectl get ns

Sample Output:

NAME              STATUS   AGE
default           Active   10d
kube-system       Active   10d
kube-public       Active   10d
kube-node-lease   Active   10d

Getting Kubernetes Resources in K3s

Displays a list of commonly used Kubernetes resources such as Pods, Nodes, Services, and Endpoints. Helps monitor resource status, troubleshoot issues, and verify deployments.

Prerequisite:

None — just ensure your cluster is running and kubectl is configured (e.g., k3s kubectl or alias).

Command Syntax:

kubectl  get <resource> -n <namespace> [flags]

Parameters & Options:

Parameter	Type	Description
`<resource>`	arg	Resource type like `pods`, `svc`, `nodes`
`-n`	flag	Specifies the namespace (where applicable)
`-o wide`	flag	Shows more details like IP, node, etc.
`-A` or `--all-namespaces`	flag	Lists resources across all namespaces

Common Patterns or Use Cases:

Get all pods in a specific namespace

kubectl  get  pods  -n  default

Get all nodes in the cluster

kubectl  get  nodes

Get all services in a namespace

kubectl  get  svc  -n  kube-system

Get endpoints in a namespace

kubectl  get  endpoints  -n  app

Get pods with extended info (e.g., IPs, node names)

kubectl  get  pods  -o  wide  -n  erpnext

Get all services across all namespaces

kubectl  get  svc  -A

Sample Output:

Pods

NAME  READY  STATUS  IP  NODE  AGE

nginx-deployment-65fdb85b49-7jdt4  1/1  Running  10.42.0.5  k3s-node1  5m

Nodes

NAME  STATUS  ROLES  AGE  VERSION

k3s-master  Ready  control-plane,master  20h  v1.27.3+k3s1

Services

NAME  TYPE  CLUSTER-IP  EXTERNAL-IP  PORT(S) AGE

kubernetes  ClusterIP  10.43.0.1 <none> 443/TCP  20h

Endpoints

NAME  ENDPOINTS  AGE

nginx  10.42.1.23:80  4m

Describing a Pod in K3s

Displays detailed information about a specific pod, including container specs, events, volume mounts, and state transitions.

Useful for debugging issues like scheduling problems, container crashes, and init delays.

Prerequisite:

You need the pod name and namespace

Command Syntax:

kubectl  describe  pod <pod-name> -n <namespace>

Parameters & Options:

Parameter	Type	Description
`<pod-name>`	arg	Name of the pod to describe
`-n`	flag	Specifies the namespace of the pod

Common Patterns or Use Cases:

Describe a pod to troubleshoot issues like CrashLoopBackOff:

kubectl  describe  pod  nginx-5d5f74c8f9-abcde  -n  web

Sample Output:

Name: nginx-5d5f74c8f9-abcde

Namespace: web

Priority: 0

Node: k3s-node/192.168.1.100

Start Time: Tue, 28 May 2025 10:21:01 +0000

Labels: app=nginx

Status: Running

IP: 10.42.0.12

Containers:

nginx:

Image: nginx:latest

Port: 80/TCP

State: Running

Ready: True

Events:

Type Reason Age From Message

---- ------ ---- ---- -------

Normal Scheduled 2m default-scheduler Successfully assigned web/nginx-xxx to k3s-node

Getting Pod Logs in K3s

Displays the logs of a specific pod’s container.

Helps identify errors, monitor application output, and debug running containers.

Prerequisite:

You need the pod name and namespace

Command Syntax:

kubectl  logs <pod-name> -n <namespace> [flags]

Parameters & Options:

Parameter	Type	Description
`<pod-name>`	arg	Name of the pod
`-n`	flag	Specifies the namespace
`-c <container>`	flag	Specify container name if multiple exist
`-f`	flag	Follow the logs output (stream logs)

Common Patterns or Use Cases:

Get logs of a pod:

kubectl  logs  myapp-pod-12345  -n  mynamespace

Follow logs live:

kubectl  logs  -f  myapp-pod-12345  -n  mynamespace

Get logs from a specific container in a pod:

kubectl  logs  myapp-pod-12345  -c  sidecar-container  -n  mynamespace

Sample Output:

2025-05-28T12:00:01.123Z INFO Starting server on port 8080

2025-05-28T12:00:05.456Z INFO Connection established with database

2025-05-28T12:05:12.789Z ERROR Failed to fetch user data

Executing Commands Inside a Pod in K3s

Allows you to run commands interactively or non-interactively inside a container of a pod.

Useful for troubleshooting, debugging, or running administrative commands inside the pod.

Prerequisite:

You need the pod name and namespace

Get pod name with:

 kubectl  get  pods  -n <namespace>

Command Syntax:

kubectl  exec  -it <pod-name> -n <namespace> -- <command>

Parameters & Options:

Parameter	Type	Description
`<pod-name>`	arg	Name of the pod
`-n`	flag	Namespace where the pod resides
`-i`	flag	Pass stdin to the container (interactive mode)
`-t`	flag	Allocate a pseudo-TTY (for terminal)
`<command>`	arg	The command to run inside the pod’s container

Common Patterns or Use Cases:

Run a single command (e.g., list files):

k3s  kubectl  exec  -it  mypod  -n  default  --  ls  /app

Run interactive bash shell:

k3s  kubectl  exec  -it  mypod  -n  default  --  /bin/bash

Sample Output:

# /bin/sh inside nginx-pod

$ ls -l /usr/share/nginx/html

total 4

-rw-r--r-- 1 root root 612 May 28 12:00 index.html

Port Forwarding to a Pod in K3s

Allows you to forward one or more local ports to a port on a pod.

Useful for debugging or accessing a pod’s service locally without exposing it via a service or ingress.

Prerequisite:

You need the pod name , namespace and the ports you want to forward.

Command Syntax:

kubectl  port-forward  pod/<pod-name> <local-port>:<pod-port> -n <namespace>

Parameters & Options:

Parameter	Type	Description
`pod/<pod-name>`	arg	Name of the pod
`<local-port>`	arg	Local machine port to listen on
`<pod-port>`	arg	Port on the pod to forward traffic to
`-n`	flag	Namespace

Common Patterns or Use Cases:

Forward local port 8080 to pod port 80:

kubectl  port-forward  pod/myapp-pod  8080:80  -n  default

Sample Output:

Forwarding from 127.0.0.1:8080 -> 80

Handling connection for 8080

Viewing Pod Performance Metrics in K3s

Displays CPU and memory usage of pods to monitor resource consumption and troubleshoot performance issues.

Prerequisite:

Metrics Server must be installed and running in the cluster.

You can verify with:

kubectl  get  deployment  metrics-server  -n  kube-system

Command Syntax:

kubectl  top  pods  -n <namespace> [flags]

Parameters & Options:

Parameter	Type	Description
`-n`	flag	Namespace to list pod metrics from
`--containers`	flag	Show metrics per container (optional)

Common Patterns or Use Cases:

Show CPU and memory usage of all pods in a namespace:

kubectl  top  pods  -n  default

Show metrics per container inside pods:

kubectl  top  pods  -n  default  --containers

Sample Output:

NAME CPU(cores) MEMORY(bytes)

nginx-5d5f74c8f9-abcde 10m 20Mi

redis-master-765d459796-xyz 5m 15Mi

Getting Pod Details in YAML Format in K3s

Outputs the full detailed configuration and status of a pod in YAML format.

Useful for deep inspection, debugging, or exporting pod specs.

Prerequisite:

You need the pod name and namespace

Command Syntax:

kubectl  get  pod <pod-name> -n <namespace> -o  yaml

Parameters & Options:

Parameter	Type	Description
`<pod-name>`	arg	Name of the pod
`-n`	flag	Namespace of the pod
`-o yaml`	flag	Output format in YAML

Common Patterns or Use Cases:

Get detailed YAML description of a pod:

kubectl  get  pod  myapp-pod  -n  default  -o  yaml

Sample Output:

apiVersion: v1

kind: Pod

metadata:

name: myapp-pod

namespace: default

labels:

app: myapp

spec:

containers:

- name: myapp-container

image: myapp:latest

ports:

- containerPort: 8080

status:

phase: Running

podIP: 10.42.1.15

Deleting Pods in K3s

Deletes one or more pods by name or label selector.

Useful for removing faulty or unneeded pods manually.

Prerequisite:

You need the pod name and namespace

Command Syntax:

kubectl  delete  pod <pod-name> -n <namespace>

Or delete multiple pods by label selector:

kubectl  delete  pod  -l <label-selector> -n <namespace>

Parameters & Options:

Parameter	Type	Description
`<pod-name>`	arg	Name of the pod to delete
`-n`	flag	Namespace of the pod
`-l <selector>`	flag	Label selector to delete matching pods

Common Patterns or Use Cases:

Delete a single pod:

kubectl  delete  pod  myapp-pod  -n  default

Delete all pods with label app=nginx:

kubectl  delete  pod  -l  app=nginx  -n  default

Sample Output:

pod "myapp-pod" deleted

Pruning Resources in Kubernetes (Kubectl)

Kubernetes can achieve pruning (cleaning up unused resources) with:

kubectl apply --prune — deletes resources not defined in your current manifests.
Manual deletion using label selectors.

Command Syntax:

kubectl  apply  -f <directory-or-file> --prune  -l <label-selector> -n <namespace>

Parameters & Options:

Parameter	Type	Description
`-f`	flag	Directory or file path containing manifests
`--prune`	flag	Enables pruning of resources not in the manifest
`-l <label>`	flag	Label selector to identify resources to prune
`-n`	flag	Namespace to apply the pruning

Common Patterns or Use Cases:

Apply manifests and prune resources without the label app=myapp in namespace default:

kubectl  apply  -f  ./manifests  --prune  -l  app=myapp  -n  default

Sample Output:

deployment.apps/myapp configured

service/myapp-service configured

pod/myapp-pod deleted

configmap/old-config deleted

Scaling Deployments in K3s

Use kubectl scale to manually increase or decrease the number of pod replicas in a deployment or other scalable resource.

Command Syntax:

kubectl  scale  deployment <deployment-name> --replicas=<count>  -n <namespace>

Parameters & Options:

Parameter	Type	Description
`<deployment-name>`	arg	Name of the deployment
`--replicas`	flag	Number of desired replicas
`-n`	flag	Namespace where the deployment is located

Common Patterns or Use Cases:

Scale a deployment to 5 replicas:

kubectl  scale  deployment  myapp-deployment  --replicas=5  -n  default

Scale down to a single replica:

kubectl  scale  deployment  myapp-deployment  --replicas=1  -n  default

Sample Output:

deployment.apps/myapp-deployment scaled

Cordoning & Uncordoning Nodes in K3s

Used to safely manage node scheduling during maintenance. Cordoning prevents new pods from being scheduled on the node, while uncordoning re-enables scheduling.

Cordon a Node

Prevents new pods from being scheduled on a specific node. Does not affect running pods.

Command Syntax:

kubectl cordon <node-name>

Parameters & Options:

Parameter	Type	Description
`<node-name>`	arg	Name of the node to cordon

Common Use Case:

kubectl cordon worker-node-1

Sample Output:

node/worker-node-1 cordoned

Uncordon a Node

Re-enables scheduling of new pods on the node.

Command Syntax:

kubectl uncordon <node-name>

Parameters & Options:

Parameter	Type	Description
`<node-name>`	arg	Name of the node to uncordon

Common Use Case:

kubectl uncordon worker-node-1

Sample Output:

node/worker-node-1 uncordoned