chaos-helm

This repo consists of helm charts for the chaos experiment custom resources. The helm installation is typically used in admin mode of litmus, wherein all chaosexperiment CRs (and other chaos resources such as chaosengine)are placed/run in a single namespace while the applications themselves are in their respective namespaces.

PreRequisites

Install helm3

Note: Litmus uses Kubernetes custom resource definitions (CRDs) to define chaos intent. Helm3 handles CRDs better than helm2!

While the order of installation is not strict, you also need to install the Litmus Chaos Operator helm chart to be able to run these experiments. Refer to these instructions to achieve the same.

Installation Steps

The following steps will help you install the generic Kubernetes chaos experiment custom resources via helm.

Step-1: Add the chaos helm repository

root@demo:~# helm repo add k8s-chaos https://litmuschaos.github.io/chaos-helm/
"k8s-chaos" has been added to your repositories

root@demo:~# helm repo list
NAME            URL                                       
k8s-chaos     https://litmuschaos.github.io/chaos-helm/

root@demo:~# helm search repo k8s-chaos
NAME                            CHART VERSION   APP VERSION     DESCRIPTION
k8s-chaos/kubernetes-chaos      1.2.0           1.2.0           A Helm chart to install litmus chaos experiment...

Step-2: Create the litmus namespace

The litmus chaos experiment CRs will be placed in this namespace.

Note: The chaos experiments can be placed in any namespace, though it is typically placed in "litmus". Please note that the chaosengine will have to be created in the same namespace.

root@demo:~# kubectl create ns litmus 
namespace/litmus created

Step-3: Install the Kubernetes chaos experiments

root@demo:~# helm install k8s k8s-chaos/kubernetes-chaos --namespace=litmus
NAME: k8s
LAST DEPLOYED: Fri Apr  3 17:44:39 2020
NAMESPACE: litmus
STATUS: deployed
REVISION: 1
TEST SUITE: None

Additional Steps (Verification)

You can run the following commands if you wish to verify if all chaos experiments are installed successfully.

Check if chaos experiments are named appropriately and available in litmus namespace

root@demo:~# kubectl get chaosexperiments -n litmus
NAME                         AGE
k8s-container-kill           13s
k8s-disk-fill                13s
k8s-disk-loss                13s
k8s-node-cpu-hog             13s
k8s-node-drain               13s
k8s-node-memory-hog          13s
k8s-pod-cpu-hog              13s
k8s-pod-delete               13s
k8s-pod-network-corruption   13s
k8s-pod-network-latency      13s
k8s-pod-network-loss         13s

Check if the chaos-admin RBAC components are installed successfully

root@demo:~# kubectl get serviceaccount -n litmus
NAME              SECRETS   AGE
default           1         11h
k8s-chaos-admin   1         90s

Check if the chaos operator RBAC is configured appropriately:

root@demo:~# kubectl get clusterrole | grep chaos-admin
k8s-chaos-admin                                        71s

root@demo:~# kubectl get clusterrolebinding | grep chaos-admin
k8s-chaos-admin                                        98s

uditgaurav / chaos-helm