Kubernetes Commands

Kubernetes basic Objects

- pod : Basic execution unit of a k8s applicaiton
- service : An abstract way to expose an app runnig on set of pods as network service.
- volume : A volume outlives any containers that run within the pod.
- namespace : Namespaces provide a scope for names

Kubernetes higher level abstractions

- deployment
- daemonset
- statefulset
- replicaset : Specified number of pod replicas are running at any one time. 
- job

Basic commands

nodes

# get the list of nodes
> kubectl get nodes

# get details of a selected node
> kubectl describe node <node name>

# delete a node
> kubectl delete node <node name>

namespace

# get the list of namespaces
> kubectl get namespaces
> kubectl get namespaces
> kubectl get ns

# create a new namespace
# > kubectl create namespace <namespace name>
> kubectl create namespace ns1

# delete a namespace
# this command will also delete all the objects under the namespace
# > kubectl delete namespace <namespace name>
> kubectl delete namespace ns1

pod

# get the list of pods running in default namespace
> kubectl get pods

# get the list of pods running in requirement namespace
> kubectl get pods -n <ns name>

# get the list of pods with wide/more options
> kubectl get pods -n <ns name> -o wide

# create a pod using pod1.yaml file
> kubectl create -f pod1.yaml

# get the details of selected pod
> kubectl describe pod <pod name>

# delete the pod from default namespace
> kubectl delete pod <pod name>

# delete the pod from required namespace
> kubectl delete pod <pod name> -n <ns name>

# get the logs of a selected pod
> kubectl logs <pod name>

# get the logs continuously of a selected pod
> kubectl logs -f <pod name>

# execute a command inside a pod
> kubectl exec -it <pod name> -- <command>

# get the terminal of a selected pod
> kubectl exec -it <pod name> -- bash

# get the terminal of a selected pod from a selected containers
# if -c is not given, the first container will execute the command
> kubectl exec -it <pod name> -c <container name> -- bash

replica sets

used to create multiple replicas of selected pod

# get list of replica-set
> kubectl get replicasets
> kubectl get replicaset
> kubectl get rs

# get details of selected replica-set
> kubectl describe rs <rs name>

# to scale out or in, update the replicas in yaml file
> kubectl apply -f <rs yaml file>

# delete a replica-set
> kubectl delete replicaset <rs name>

service

used to balance the load amongst multiple pods
these multiple pods can be created using replica-set or deployment
types
ClusterIP
- service which will be accessible only within the cluster
- service can not be accessed outside the cluster
- can be used to access an application inside the cluster by other pods
- e.g. frontend pod is accessing backend service which is load balancing the backend pods
- ports
- port
  - the internal client will send the request to service on this port
  - you are free to choose this port as per your requirement
- targetPort
  - service will forward the request to pod(s) on this port
  - this port number must be same as the port on which the pod is listening on
NodePort
- service will make the application accessible outside the cluster
- it internally will create a clusterIP service
- ports
- port
  - the internal client will send the request to service on this port
  - you are free to choose this port as per your requirement
- targetPort
  - service will forward the request to pod(s) on this port
  - this port number must be same as the port on which the pod is listening on
- nodePort
  - the port assigned to the node on which external client will send the request
  - if needed you can specify the nodePort within the range of 30000-32767
  - if not specified, the kubernetes will assign a random nodePort to the service
LoadBalancer
- used to create a load balancer in cloud (for AWS it will create ALB)

# get the list of services
> kubectl get services

# get the service details
> kubectl describe service <service name>

config map

collection of key-value pairs (configuration)
used for storing non-sensitive application configurations
e.g. port number, backend url
all the configurations stored in config map are exposed to the application via environment variables
all values must be in string format (wrapped in double quotes)

# get the list of config maps
> kubectl get configmap
> kubectl get cm

# get details of selected config map
> kubectl describe cm <cm name>

# delete selected config map
> kubectl delete cm <cm name>

secrets

collection of key-value pairs (configuration)
used for storing sensitive application configurations
e.g. password, secret, access token
all the configurations stored in secrets are exposed to the application via environment variables
all values must be in bas64 encoded string format (wrapped in double quotes)

# get the list of secrets
> kubectl get secrets

# get details of a selected secret
> kubectl describe secret <secret name>

# delete selected secret
> kubectl delete secret <secret name>

deployment

represents logical deployment of an application
internally it uses replica set to replicate the pods
can be updated or rollbacked using rollout commands

# get the list of deployments
> kubectl get deployments
> kubectl get deploy

# get details of selected deployment
> kubectl describe deploy <deploy name>

# delete deployment
> kubectl delete deploy <deploy name>

rollout

# restart the deployment using rollout
# this will force deployment to load the new version from docker hub
> kubectl rollout restart deployment <deployment-name>

# get the history of rollout
> kubectl rollout history deployment <deployment-name>

# rollback to the older version (previous version)
> kubectl rollout undo deployment <deployment-name>

# rollback to the specific older version
> kubectl rollout undo deployment <deployment-name> --to-revision=<version-number>

# update the image tag (version)
> kubectl set image deployment <deployment-name> <container-name>=<newer version>

# get the current status of rollout
> kubectl rollout status deployment <deployment-name>

persistent volumes

# get the list of persistent volumes
> kubectl get persistentvolumes
> kubectl get pv

# create a pv
> kubectl apply -f pv.yaml

# get details of selected pv
> kubectl describe pv <pv name>

# delete a pv
> kubectl delete pv <pv name>

persistent volume claim

# get the list of pvc
> kubectl get pvc

# get details of a selected pvc
> kubectl describe pvc <pvc name>

# delete a pvc
> kubectl delete pvc <pvc name>

metrics service

# apply the metrics server yaml
> kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

# the above url will deploy the metrics server in kube-system namespace
# by default, this will not work
# to fix the problem
> kubectl edit deployment metrics-server -n kube-system
# add the following line on line number 45
> --kubelet-insecure-tls=true
# save and exit => esc :wq

# get the top node usage
> kubectl top nodes

# get the top pods usage
> kubectl top pods

# get all the resources created in the kube-system namespace
> kubectl get all -n kube-system

horizontal pod autoscaling

# get the list of hpa
> kubectl get hpa

# get details of selected hpa
> kubectl describe hpa <hpa name>

# delete a selected hpa
> kubectl delete hpa <hpa name>

job

# get the list of jobs
> kubectl get jobs

# get details of selected job
> kubectl describe job <job name>

# delete a selected job
> kubectl delete job <job name>

cron job

# get the list of cronjobs
> kubectl get cronjobs

# get details of selected cronjob
> kubectl describe cronjob <cronjob name>

# delete a selected cronjob
> kubectl delete cronjob <cronjob name>

Back to CheatSheets