Common kubectl Commands

The following commands are executed on the bmc node running the K3s Server. Because the K3s Server runs in a Docker container, all commands use the following prefix:

sudo docker exec k3s-server kubectl

Cluster Information

Pod Management

Service Management

A Service provides a stable access entry point for a set of Pods. After a Pod is created, deleted, or restarted, its IP address may change, so other applications should not depend on Pod IPs directly. A Service locates the corresponding Pods through a label selector and forwards requests to the available Pods.

In short:

• Pod: The container instance that actually runs the application.
• Service: Provides a stable name and access address for the Pods.
• Client: Accesses the application through the Service without needing to know the specific Pod IPs.

The common Service types are as follows:

In the current K3s environment, you can use NodePort to expose a web service outside the cluster. For example, after a Service maps container port 80 to a port on the node, the application can be accessed at http://<node-ip>:<NodePort>.

Deployment Management

A Deployment manages applications that need to run continuously, such as web services, API services, and background services. It maintains a specified number of Pod replicas through a ReplicaSet and supports rolling updates, automatic fault recovery, and replica scaling.

Applicable scenarios:

• Long-running stateless services
• Applications that require multiple Pod replicas
• Applications that need rolling updates or fast rollbacks

Job Management

A Job runs one-off or time-limited tasks, such as data processing, initialization operations, and batch tasks. A Job creates Pods and keeps retrying until the task completes successfully or the configured number of retries is reached.

Applicable scenarios:

• One-off data processing
• Database initialization or migration
• Batch file processing
• Cluster or application initialization tasks

CronJob Management

A CronJob periodically creates Jobs according to a preset schedule, for example scheduled backups, log cleanup, and periodic data synchronization. Each scheduled run creates a Job, which in turn creates and runs a Pod.

Applicable scenarios:

• Scheduled backups
• Periodic cleanup of temporary files or logs
• Periodic data synchronization
• Scheduled report generation

A CronJob uses standard cron expressions to set the execution time. For example, */5 * * * * means running every 5 minutes.

Namespaces and Configuration

Namespaces divide mutually isolated resource scopes within the same K3s cluster. You can create different namespaces by project, team, or environment — for example dev, test, and prod — to avoid resource name conflicts between applications and to make permission and resource management easier.

ConfigMaps and Secrets store the configuration required by applications:

• ConfigMap: Stores non-sensitive configuration such as environment variables, configuration files, and service addresses.
• Secret: Stores sensitive information such as passwords, tokens, and certificates. A Secret is not a complete encrypted storage solution; you should still restrict access and avoid committing sensitive content to code repositories.

When creating an application, you can specify the namespace in the resource manifest via metadata.namespace; when running commands, you can specify the target namespace with -n <namespace>. If no namespace is specified, commands apply to the current namespace, which is usually default.

Resources and Troubleshooting