25 Kubernetes questions that will help you get certified and become more efficient with K8s

Export the pod to a yaml definition using the following command:

kubectl get pod yourpod -o yaml > pod.yaml

Create a new deployment definition using the following command:

kubectl create deployment deploymentname --image=imagename --dry-run=client -o yaml > deployment.yaml

Display the pod yaml:

cat pod.yaml

Copy the spec part

Edit the deployment file:

nano deployment.yaml

Paste the pod .spec block to deployment template .spec, correct indentation, save the file.

Create the deployment using the file:

kubectl apply -f deployment.yaml

Run the following command:

kubectl expose pod pod1 --name=svc1 --port=8080 --target-port=80

Run the following command:

kubectl run tmp --image=busybox --rm -i -- /bin/sh -c "wget -m 5 svc1:80"

The -m 5 parameter defines maximum time allowed for the transfer. It is used as a 5 second timeout. If the response is not transferred by that time, wget stops.

• Imperative commands
• Imperative object configuration
• Declarative object configuration

Very often you generate a manifest using an imperative command together with --dry-run=client parameter and then further tweak it using an editor. Once finished, you apply the changes.

Example:

Run the following to generate a pod manifest.

kubectl run mypod --image=nginx --dry-run=client -o yaml > file.yaml

Run the following to open the pod manifest in an editor:

nano file.yaml

Apply the changes by running the following command:

kubectl apply -f file.yaml

The above technique will also allow you to get back to the manifest if needed. You can always delete the resoruce (pod), modify the manifest and recreate it again using the modified version.

Run the following command:

kubectl get pod --watch

Run one of the following commands (they are all the same):

kubectl get sa

kubectl get serviceaccount

kubectl get serviceaccounts

Run the following command:

kubectl cluster-info

Configure alias, so that you don't need to type kubectl.

Run the following command:

alias k=kubectl

Then you can type kubectl commands like this:

k get pod

Option 1:

Create alias and variable:

alias k=kubectl
    export n="config set-context --current --namespace"

Then switch namespace like this:

k $n=default

Option 2:

Create two aliases

alias k=kubectl
alias kn="kubectl config set-context --current --namespace"

Then switch namespace like this:

kn default

Run the following command:

kubectl top pod --sort-by="cpu" -o wide

You can also disable headers by using --no-headers parameter.
To display information about usage of containers, use --containers=true parameter.

Cordon marks the Kubernetes node unschedulable to prevent new pods from arriving. This is usually done as a preparation for node upgrade.

To cordon the node, run the following command:

kubectl cordon nodename

Uncordon makes the node schedulable again after it has been cordoned before. This command is usually used after the node has been upgraded to higher version (for which it needed to be cordoned).

To uncordon the node, run the following command:

kubectl uncordon nodename

Drain marks the node as unschedulable to prevent new pods from arriving. It also evicts the pods from the node.

Drain is executed to drain the node in preparation for maintenance.

To drain the node, run the following command:

kubectl drain nodename

To get IP addresses of all nodes, run the following command:

kubectl get nodes -o wide

Option 1:

Run the following command:

kubectl config current-context

Option 2:

Run the following command:

kubectl config get-contexts|grep "*"

Run the following command:

kubectl config get-contexts -o name

Use the following command:

cat ~/.kube/config|yq eval -o json|jq ".contexts[].name"

How does it work?

• The configuration data is stored the file ~/.kube/config
• cat command displays the conent of the file
• yq command converts the output from yaml to JSON format
• jq command will iterate over all child elements of element (array) contexts, name property of each child element is displayed.

Run the following command:

kubectl get all

Run the following command:

kubectl get deploy,po,svc -o yaml

Run the following command:

kubectl get node -o yaml|grep --ignore-case --context=10 ip

How it works?

• Using kubectl, details of nodes are displayed in yaml format
• Using grep command, rows are searched for string ip (case insensitive search), besides the row, block of 10 rows before and 10 rows after is displayed
• This allows you to quickly find key blocks in manifests when you don't know exactly what you are looking for. Based on the output you can further refine the query.

Run the following command:

kubectl get po pod1 -o json

Run the following command to find a master node:

kubectl get node|grep master

Copy the node name and run the following command:

kubectl get pod -o wide|grep your-master-node-name

Option 1:

Create a deamonset.

Option 2:

Create a deployment that has pods with antiaffinity that allows only one pod to run on a node. Make sure the count of replicas is the same or higher than the number of nodes.

Run the following command

kubectl get ep service1