Wiki

* Scroll me *

PREFACE

About Wiki

GOLANG

SHELL

NETWORKING

KUBERNETES

ISTIO

VAULT

INFRA AS CODE

Kubernetes Commands

API resources

List all API objects in a namespace

kubectl api-resources --verbs=list --namespaced -o name \
  | xargs -n 1 kubectl get --show-kind --ignore-not-found -n ${NAMESPACE}

because kubectl get all does not really give you all the resources.

Events

List events in the last hours and sort by time.

kubectl get events --sort-by='.lastTimestamp' -A

Traffic

Port-forwarding

Listen on port 8888 locally, forwarding connection to 5000 on ${POD_NAME}

kubectl -n ${NAMESPACE} port-forward pod/${POD_NAME} 8888:5000

Note that pod/mypod can also be svc/ or deployment/.

This is useful to troubleshoot in-cluster services without exposing them. Port-forwarding works only for TCP traffic at the moment.

Pods and Shells

Execute a command in a Pod

kubectl exec ${POD_NAME} -- echo hello world

kubectl exec ${POD_NAME} -c ${CONTAINER_NAME} -- curl https://www.example.org

By label

kubectl exec -it \
  $(kubectl get pod -n ${NAMESPACE} -l key=value --output=jsonpath={.items..metadata.name}) \
  -n ${NAMESPACE} -c ${CONTAINER_NAME} -- bash

Open a shell to a Pod

kubectl exec -it ${POD_NAME} -- bash

Spin up a debug Pod and open a shell to it

kubectl run -it ${POD_NAME} --image=debian --rm --command -- sh

This is useful when the application Pod is super stripped down, such as using distroless base image that makes troubleshooting difficult due to the lack to tooling.

Find all unhealthy Pods

kubectl get po -A | grep -v Running | grep -v Completed

Get Pod name by label

kubectl get po -n ${NAMESPACE} -l key=value --output=jsonpath={.items..metadata.name}

Logs

Get last 20 lines of logs

kubectl logs --tail=20 ${POD_NAME}

Get logs in the last 3 hours

kubectl logs --since=3h ${POD_NAME}

Get tail logs and stream new logs

kubectl logs --tail=20 ${POD_NAME} -f

####Aggregate logs from multiple Pods using label selector

kubectl logs -l app=server --tail=20 -f

Nodes

List all used NodePorts in a cluster

$ TEMPLATE='{{range .items}}{{range.spec.ports}}{{if .nodePort}}{{.nodePort}}{{.}}{{"\n"}}{{end}}{{end}}{{end}}'
$ kubectl get svc --all-namespaces -o go-template="${TEMPLATE}"

List all Nodes by creation time

kubectl get node --sort-by=.metadata.creationTimestamp

Get Pod Count per Node

for n in $(kubectl get nodes --no-headers | cut -d " " -f1); do
    echo -n "${n}: "; kubectl get pods --all-namespaces  --no-headers --field-selector spec.nodeName=${n} | wc -l
done

Useful kubectl plugins

Namespace-wide rolling restart

Repo: LifeWay/kubectl-roll-plugin

kubectl roll -n ${NAMEPSACE} will trigger a rolling restart of all StatefulSets, DaemonSets, and Deployments in a given namespace

Charles Xu