What is the concept of Pod in Kubernetes? What are its lifecycle and restart policies?

A Pod in Kubernetes is the smallest deployable unit that contains one or more closely related containers. These containers share network and storage resources.

Pod Characteristics

1. Shared Network Namespace: Containers in the same Pod share the same IP address and port space, and can communicate with each other via localhost.

2. Shared Storage Volumes: Containers in a Pod can share mounted storage volumes, enabling data sharing and persistence.

3. Atomic Scheduling: The Pod is scheduled as a whole to run on the same Node.

4. Ephemeral: Pods are temporary and replaceable. When a Pod is deleted or a Node fails, the Pod does not automatically recover.

Pod Lifecycle

The Pod lifecycle includes the following phases:

1. Pending: The Pod has been created, but the containers have not started yet, possibly because the image is downloading or resources are insufficient.

2. Running: All containers in the Pod have been created, and at least one container is running.

3. Succeeded: All containers in the Pod have terminated successfully.

4. Failed: All containers in the Pod have terminated, but at least one container terminated in a failed state.

5. Unknown: Unable to get the Pod's status, usually because communication with the Node where the Pod is located has failed.

Pod Restart Policies

Kubernetes supports three Pod restart policies:

1. Always: Always restart the container when it fails. This is the default policy.

2. OnFailure: Only restart when the container fails with a non-zero exit code.

3. Never: Do not restart when the container fails.

Pod and Container Relationship

A Pod is a wrapper for containers. A Pod can contain:

• A single main container (most common)
• One main container plus one or more sidecar containers (Sidecar pattern)
• Multiple cooperating containers

Best Practices

1. One Pod One Container: For most applications, it is recommended that a Pod contains only one container, which makes management and scaling easier.

2. Use Sidecar Pattern: When multiple closely cooperating containers are needed, you can use the Sidecar pattern, such as log collection, monitoring agents, etc.

3. Avoid Running Multiple Unrelated Containers in a Pod: This increases management complexity and is not conducive to scaling and troubleshooting.

4. Set Resource Limits Reasonably: Set CPU and memory requests and limits for Pods to avoid resource contention.

5. Use Health Checks: Configure livenessProbe and readinessProbe to ensure the health status of the Pod.