Essential Guide for Beginners

Kubernetes provides a series of commands to assist in debugging and pinpointing issues. This section lists some common commands to help application administrators troubleshoot and resolve issues quickly.

1. Identifying Problems

Before addressing any issues, it’s crucial to determine the type of problem: whether it involves a Pod, Service, or Controller (e.g., Deployment, StatefulSet). Use appropriate commands based on the identified issue to investigate the root cause.

2. Common Pod Commands

If a Pod is in a Pending state, crashes repeatedly, or cannot accept traffic, the following commands can help identify the underlying issue:

1. Get Pod status

kubectl -n ${NAMESPACE} get pod  -o wide

2. View the yaml configuration of the Pod

kubectl -n ${NAMESPACE} get pod ${POD_NAME}  -o yaml

3. View Pod events

kubectl  -n ${NAMESPACE} describe pod ${POD_NAME}

4. View Pod logs

kubectl  -n ${NAMESPACE} logs ${POD_NAME} ${CONTAINER_NAME}

5. Log into Pod

kubectl -n ${NAMESPACE} exec -it  ${POD_NAME} /bin/bash

3. Common Controller Commands

The Controller is responsible for Pod’s lifecycle management. If a Pod fails to register, the Controller commands can help identify the cause. Here’s how to use common commands for a Deployment as an example:

1. View Deployment status

kubectl -n ${NAMESPACE} get deploy -o wide

2. See Deployment yaml configuration

kubectl -n ${NAMESPACE} get deploy ${DEPLOYMENT_NAME} -o yaml

3. View Deployment events

kubectl -n ${NAMESPACE} describe deployment ${DEPLOYMENT_NAME}

4. Common Service Commands

Services define how to access a set of Pods. If the application is inaccessible, use the following Service commands to diagnose the issue:

1. View Service status

kubectl  -n ${NAMESPACE} get svc -o wide

This command provides details about the Service type, internal and external IPs, exposed ports, and Selector information.

2. View Service events and load balancing information

kubectl  -n ${NAMESPACE} describe svc ${SERVICE_NAME} 
 
Name:              example-app
Namespace:         default
Labels:            app=example-app
Annotations:       <none>
Selector:          app=example-app
Type:              ClusterIP
IP:                10.2.192.27
Port:              web  8080/TCP
TargetPort:        8080/TCP
Endpoints:         192.168.59.207:8080,192.168.75.87:8080,192.168.84.90:8080
Session Affinity:  None
Events:            <none>

By using this command, you can check the Endpoints information of the Service. If the Endpoints are empty, it indicates a configuration error, preventing the Service from forwarding traffic to the corresponding Pods. Ensure that the Port and TargetPort settings align with the actual exposed ports of your application.