node-drain

Cordons a node (marks it unschedulable) and evicts all eligible pods. Simulates a planned maintenance event.

Target kind: Node
Implementation: Kubernetes API (no daemon)
Rollback: Yes — uncordons the node

Parameters

Name	Type	Required	Default	Description
timeout	string	No	"5m"	Max time to wait for pod eviction
ignoreDaemonSets	string	No	"true"	Skip DaemonSet pods during eviction
deleteEmptyDirData	string	No	"true"	Evict pods with emptyDir volumes

Example

apiVersion: chaos.chaosplane.io/v1alpha1
kind: ChaosExperiment
metadata:
  name: node-drain-example
  namespace: chaosplane
spec:
  target:
    kind: Node
    labelSelector:
      matchLabels:
        node-role.kubernetes.io/worker: ""
  action:
    type: node-drain
    parameters:
      timeout: "5m"
      ignoreDaemonSets: "true"
      deleteEmptyDirData: "true"
  duration: 5m
  rollback:
    enabled: true

Rollback behavior

Sets spec.unschedulable: false on the node via the Kubernetes API (uncordon). Pods are not automatically rescheduled back to the node — the scheduler will place new pods there once it's uncordoned.

Implementation notes

Uses the Kubernetes API directly from the operator. No daemon involvement. The operator:

1. Sets node.Spec.Unschedulable = true (cordon)
2. Lists all pods on the node
3. Skips mirror pods (static pods with kubernetes.io/config.mirror annotation)
4. Skips DaemonSet pods if ignoreDaemonSets: "true"
5. Skips pods with emptyDir volumes if deleteEmptyDirData: "false"
6. Evicts remaining pods via PolicyV1().Evictions().Evict()

The eviction respects PodDisruptionBudgets.