set -o vi
alias e=vim
alias c=clear
alias kg="k get"
alias kd="k describe"
alias kr="k run"
alias ke="k exec"
alias ka="k apply -f"
alias kc="k config"
alias kx="kc use-context" # poor mans kubectx
alias kn="kc set-context --current --namespace" # poor mans kubens
del() { k delete "$@" --grace-period 0 --force; } # delete a resource without waiting
run() { kr tmp --image nginx:alpine --rm -it -- "$@"; del pod tmp; } # run a cmd in an existing pod
s() { ssh ckad$1; } # ssh into ckad scenario
p() { scp .bash_aliases ckad$1:~/; } # scp custom aliases into remote
export VIMINIT="set ai" # set autoindent to work on remote
| expression |
meaning |
example |
!! |
last command |
sudo !! |
!$ |
last arg of prev cmd |
vim !$ |
!* |
all args of prev cmd |
cp !* ~/backup/ |
!^ |
first arg of prev cmd |
cat !^ |
!:2 |
2nd word (0=cmd, 1=first arg…) |
echo !:2 |
!:2-4 |
words 2 through 4 |
cp !:2-4 /tmp/ |
!:2* |
word 2 to end |
rm !:2* |
!cmd |
last cmd starting with "cmd" |
!git |
!?str? |
last cmd containing "str" |
!?kubectl? |
^old^new |
quick substitution in last cmd |
^typo^fixed |
!!:gs/a/b |
global substitution on last cmd |
!!:gs/dev/prd |
| expression |
meaning |
example |
"${@:2}" |
all args from $2 onward |
f() { cmd "${@:2}"; } |
"${@:2:3}" |
3 args starting at $2 |
f() { cmd "${@:2:3}"; } |
"${@:(-2)}" |
last 2 args |
f() { cmd "${@:(-2)}"; } |
"${@:1:$#-1}" |
all args except last |
f() { cmd "${@:1:$#-1}"; } |
$# |
arg count |
[ $# -lt 2 ] && exit 1 |
$0 |
script / function name |
echo "Usage: $0 <file>" |
kn <ns> # switch namespace — do FIRST on every question
kx <ctx> # switch context - rarely needed
kc view --minify # verify current context - if unsure
# Generate base yaml
kr pod1 --image=nginx $do > pod.yaml
kr pod1 --image=busybox $do --command -- sh -c "sleep 2d" > pod.yaml
e pod.yaml
ka pod.yaml
kg pod
kd pod
k logs pod1
k logs pod1 -c container-name # multi-container pod
ke -it pod1 -- sh
ke -it deploy/frontend -- sh
ke deploy/frontend -- wget api:2222
del pod1
# Temporary pod for testing
run -- curl api:2222
run -- wget www.google.com
k create deploy nginx --image=nginx --replicas=3
k scale deploy nginx --replicas=5
k set image deploy/nginx nginx=nginx:1.25
k rollout status deploy/nginx
k rollout history deploy/nginx
k rollout history deploy/nginx --revision=3 # inspect specific revision image
k rollout undo deploy/nginx # undo to previous revision
k rollout undo deploy/nginx --to-revision=2
k edit deploy nginx # live edit, saves immediately
# Blue/Green — two full deployments, switch service selector
# 1. deploy v2 alongside v1 (different label e.g. version: v2)
# 2. k edit svc my-svc → change selector to point to v2
# 3. del old deployment when confirmed
# Canary — route small % of traffic to new version
# 1. deploy v2 with fewer replicas (e.g. 1 vs 9 for v1)
# 2. both deployments share same service selector label
# 3. gradually scale v2 up, v1 down
k create cronjob backup --schedule="*/1 * * * *" --image=busybox -- sh -c 'date'
| Position |
Field |
Range |
| 1 |
Minute |
0–59 |
| 2 |
Hour |
0–23 |
| 3 |
Day of Month |
1–31 |
| 4 |
Month |
1–12 |
| 5 |
Day of Week |
0–6 |
| Char |
Meaning |
Example |
* |
Every value |
* |
, |
List |
1,5 |
- |
Range |
1-5 |
/ |
Step |
*/10 |
| Macro |
Equivalent |
@yearly / @annually |
0 0 1 1 * |
@monthly |
0 0 1 * * |
@weekly |
0 0 * * 0 |
@daily / @midnight |
0 0 * * * |
@hourly |
0 * * * * |
# ConfigMap
k create cm app-config --from-literal=KEY=val
k create cm web-config --from-file=index.html=/path/to/file # key=filename
# Secret
k create secret generic app-secret --from-literal=user=test --from-literal=pass=pwd
kg pod -l app=v1
k label pod nginx app=v2 --overwrite
k annotate pod nginx description='desc'
# Bulk operations — modifies ALL matching pods in one shot
k label pod -l type=worker protected=true
k label pod -l type=runner protected=true
k annotate pod -l protected=true protected="do not delete this pod"
k create sa my-sa
k create role pod-reader --verb=get,list,watch --resource=pods
k create rolebinding rb --role=pod-reader --serviceaccount=default:my-sa
k auth can-i list pods --as=system:serviceaccount:default:my-sa
# ClusterRole / ClusterRoleBinding for cluster-wide access
k create clusterrole pod-reader --verb=get,list,watch --resource=pods
k create clusterrolebinding crb --clusterrole=pod-reader --serviceaccount=default:my-sa
# Expose pod or deployment
k expose pod nginx --port=80 --name=nginx-svc
k expose deploy nginx --port=80 --name=nginx-svc
k expose pod my-pod --name=my-svc --port=3333 --target-port=80
# Service types:
# ClusterIP — internal only (default)
# NodePort — exposes on every node IP:port (external, manual)
# LoadBalancer — cloud LB, includes NodePort (external, automatic)
# NodePort fields (outside→inside):
# nodePort: 30100 — external traffic hits node on this port
# port: 8080 — service listens on this port inside cluster
# targetPort: 80 — forwards to container on this port
# Fix selector mismatch (common exam issue):
kg svc my-svc -o yaml | grep -A3 selector # check svc selector
kg pods --show-labels # check pod labels
k edit svc my-svc # fix selector to match pods
# DNS:
# within same namespace: curl api:2222
# cross-namespace: curl api.venus:2222
kg endpoints my-svc
k create ingress my-ing --rule="host/path=svc:port"
kg ingress
kd ingress my-ing
k logs pod1
k logs pod1 --previous # logs from crashed container
k logs pod1 -c container-name # specific container
k logs deploy/my-deploy -c sidecar-con # sidecar logs
k top pod # resource usage
k top pod --containers # per container
k top node
kd pod nginx
kd pod nginx | grep -A5 Events # most useful — shows what's wrong
k logs pod1 --previous # logs from crashed container
kg pod nginx -o yaml > fix.yaml
e fix.yaml
del pod nginx && ka fix.yaml # force recreate
kg events -n my-ns # cluster-wide events
k top pod # check resource usage
k explain pod.spec.containers
k explain job.spec
k explain networkpolicy.spec.egress
kg pod -o jsonpath='{.status.podIP}'
kg pods -o custom-columns=NAME:.metadata.name,IP:.status.podIP
kg pods --show-labels
kg all -n my-ns # get everything in namespace
k api-resources # list all resource types with shortnames
k api-resources --namespaced=true # only namespaced resources# run sudo first
sudo -i
# docker
docker build -t registry.example.com/image:tag /path
docker push registry.example.com/image:tag
docker ps
# podman
podman images
podman run -d --name my-container registry.example.com/image:tag
podman logs my-container > /path/to/logs
helm repo add <n> <url>
helm repo update
helm search repo <chart> --versions # find available versions
helm install <release> <chart>
helm install <release> <chart> --set replicaCount=2
helm upgrade <release> <chart> # upgrade to latest
helm delete <release>
helm list # list all releases
helm list --pending # find stuck/broken releases