Kubernetes Pod Controllers

Pod controllers are essential components in Kubernetes that ensure your application workloads are running as expected. They manage the lifecycle of pods, providing features like replication, scaling, self-healing, and rolling updates. Here’s an overview of the key types of pod controllers:

  1. ReplicationController
  2. ReplicaSet
  3. Deployment
  4. DaemonSet
  5. StatefulSet

These controllers are consistent across both open-source Kubernetes and managed Kubernetes environments (AKS, EKS, GKE), as they are part of the Kubernetes software itself.

Understanding Labels and Selectors

Kubernetes is often about integrating services with each other, which is achieved through labels and selectors. Labels are key/value pairs attached to objects, such as pods, to identify and group them. Selectors enable controllers to manage pods based on their labels.

ReplicationController

  • Purpose: Ensures a specified number of pod replicas are running at any given time.
  • Functionality: Manages all pods with labels matching its selector.
    • If a pod is deleted or damaged, the ReplicationController will recreate it.
    • During deployment, if existing pods with the same labels are found, it will start managing them.
    • If no pods with matching labels are found, it will create new pods.

Example configuration:

apiVersion: v1
kind: ReplicationController
metadata:
name: my-replication-controller
labels:
app: my-app
spec:
replicas: 3
selector:
matchLabels:
app: my-app
template:
metadata:
labels:
app: my-app
spec:
containers:
- name: my-container
image: my-app-image:latest
ports:
- containerPort: 80

ReplicaSet

  • Purpose: Similar to ReplicationController but with more flexible and expressive pod selection criteria.
  • Functionality: Can manage pods with different labels using matchExpressions.

Difference between ReplicationController and ReplicaSet

  • ReplicationController: Manages all pods with labels that match the selector.
  • ReplicaSet: Can manage pods with different labels using matchExpressions and matchLabels.

Deployment

  • Purpose: Manages stateless applications, providing declarative updates.
  • Key Features: Rolling updates, rollback, scaling, and self-healing.
  • Functionality: Works on top of ReplicaSet, managing pods for high availability and ensuring smooth updates without downtime.
    • Rolling Updates: Incrementally replaces old pods with new ones.
    • Rollback: Allows easy rollback to a previous version if something goes wrong.

DaemonSet

  • Purpose: Ensures a copy of a specific pod runs on all (or some) nodes.
  • Functionality: Useful for deploying system-level and cluster-wide services, such as monitoring agents and log collectors, ensuring consistent service across the entire cluster.

StatefulSet

  • Purpose: Manages stateful applications requiring stable, unique network identifiers, persistent storage, and ordered deployment and scaling.
  • Example Use Case: Active/Passive Database Setup
    • Active Database Pod: Handles all read, write, and delete requests, acting as the primary instance.
    • Passive Database Pod: Sits idle, monitors the active pod, and takes over if the active pod fails, ensuring data consistency through synchronization.

Summary

Pod controllers in Kubernetes play a crucial role in managing the lifecycle, availability, and scalability of your application workloads. They ensure that your applications run reliably, handle updates smoothly, and maintain the desired state even in the face of failures. Understanding these controllers and how they leverage labels and selectors is essential for effective Kubernetes cluster management.