|Type||Description||Tested K8s Platform|
|Generic||Exhaust Memory resources on the Kubernetes Node||GKE, EKS|
- Ensure that the Litmus Chaos Operator is running by executing
kubectl get podsin operator namespace (typically,
litmus). If not, install from here
- Ensure that the
node-memory-hogexperiment resource is available in the cluster by executing
kubectl get chaosexperimentsin the desired namespace. If not, install from here
- There should be administrative access to the platform on which the Kubernetes cluster is hosted, as the recovery of the affected node could be manual. For example, gcloud access to the GKE project
- Application pods are healthy on the respective Nodes before chaos injection
- Application pods may or may not be healthy post chaos injection
- This experiment causes Memory resource exhaustion on the Kubernetes node. The experiment aims to verify resiliency of applications whose replicas may be evicted on account on nodes turning unschedulable (Not Ready) due to lack of Memory resources.
- The Memory chaos is injected using a job running the linux stress-ng tool (a workload generator). The chaos is effected for a period equalling the TOTAL_CHAOS_DURATION and upto MEMORY_PERCENTAGE(out of 100).
- Application implies services. Can be reframed as: Tests application resiliency upon replica evictions caused due to lack of Memory resources
- Node Memory Hog can be effected using the chaos library:
- The desired chaos library can be selected by setting
litmusas value for the env variable
Steps to Execute the Chaos Experiment
This Chaos Experiment can be triggered by creating a ChaosEngine resource on the cluster. To understand the values to provide in a ChaosEngine specification, refer Getting Started
Follow the steps in the sections below to create the chaosServiceAccount, prepare the ChaosEngine & execute the experiment.
- Use this sample RBAC manifest to create a chaosServiceAccount in the desired (app) namespace. This example consists of the minimum necessary role permissions to execute the experiment.
Sample Rbac Manifest
apiVersion: v1 kind: ServiceAccount metadata: name: node-memory-hog-sa namespace: default labels: name: node-memory-hog-sa apiVersion: rbac.authorization.k8s.io/v1beta1 kind: ClusterRole metadata: name: node-memory-hog-sa labels: name: node-memory-hog-sa rules: - apiGroups: ["","litmuschaos.io","batch","apps"] resources: ["pods","jobs","pods/log","events","chaosengines","chaosexperiments","chaosresults"] verbs: ["create","list","get","patch","update","delete"] - apiGroups: [""] resources: ["nodes"] verbs: ["get","list"] apiVersion: rbac.authorization.k8s.io/v1beta1 kind: ClusterRoleBinding metadata: name: node-memory-hog-sa labels: name: node-memory-hog-sa roleRef: apiGroup: rbac.authorization.k8s.io kind: ClusterRole name: node-memory-hog-sa subjects: - kind: ServiceAccount name: node-memory-hog-sa namespace: default
- Provide the application info in
- Provide the auxiliary applications info (ns & labels) in
- Override the experiment tunables if desired in
- To understand the values to provided in a ChaosEngine specification, refer ChaosEngine Concepts
Supported Experiment Tunables
|Variables||Description||Specify In ChaosEngine||Notes|
|TOTAL_CHAOS_DURATION||The time duration for chaos insertion (seconds)||Optional||Defaults to 120|
|LIB||The chaos lib used to inject the chaos||Optional||Defaults to `litmus`|
|MEMORY_PERCENTAGE||The size as percent of total available memory||Optional||Defaults to 90|
|RAMP_TIME||Period to wait before and after injection of chaos in sec||Optional|
|INSTANCE_ID||A user-defined string that holds metadata/info about current run/instance of chaos. Ex: 04-05-2020-9-00. This string is appended as suffix in the chaosresult CR name.||Optional||Ensure that the overall length of the chaosresult CR is still < 64 characters|
Sample ChaosEngine Manifest
apiVersion: litmuschaos.io/v1alpha1 kind: ChaosEngine metadata: name: nginx-chaos namespace: default spec: # It can be true/false annotationCheck: 'false' # It can be active/stop engineState: 'active' #ex. values: ns1:name=percona,ns2:run=nginx auxiliaryAppInfo: '' appinfo: appns: 'default' applabel: 'app=nginx' appkind: 'deployment' chaosServiceAccount: node-memory-hog-sa monitoring: false # It can be delete/retain jobCleanUpPolicy: 'delete' experiments: - name: node-memory-hog spec: components: env: # set chaos duration (in sec) as desired - name: TOTAL_CHAOS_DURATION value: '120' ## specify the size as percent of total available memory (in percentage %) ## default value 90% - name: MEMORY_PERCENTAGE value: '90'
Create the ChaosEngine Resource
Create the ChaosEngine manifest prepared in the previous step to trigger the Chaos.
kubectl apply -f chaosengine.yml
Watch Chaos progress
Setting up a watch of the Memory consumed by nodes in the Kubernetes Cluster
watch kubectl top nodes
Check Chaos Experiment Result
Check whether the application is resilient to the memory hog, once the experiment (job) is completed. The ChaosResult resource name is derived like this:
kubectl describe chaosresult nginx-chaos-node-memory-hog -n <application-namespace>
Node Memory Hog Demo [TODO]
- A sample recording of this experiment execution is provided here.