Open rudpot opened 1 year ago
Also tried using kube-aws
. That doesn't barf but also doesn't start any resources when checking with kubectl get po -n litmus
I won't venture an opinion on where the problem exists - it originates in CDK but really litmus should be validating the resource names it's trying to generate. Workaround until that happens: you can set "release" in CDK:
// Install Litmus via helm chart
// https://chaos-mesh.org/docs/production-installation-using-helm/
eksCluster.addHelmChart('Litmus', {
chart: 'litmus',
repository: 'https://litmuschaos.github.io/litmus-helm/',
namespace: 'litmus',
// Override auto-generated hash from kubectl-provider to avoid name length issues
release: 'litmus',
});
Possibly related to #157
Trying to install litmus in AWS EKS using AWS CDK fails with the following error message:
This seems to be caused by https://github.com/litmuschaos/litmus-helm/blob/master/charts/litmus/templates/server-cluster-role.yaml#L44 where the role name is expanded from
name: litmus-cluster-scope-for-{{ include "litmus-portal.fullname" . }}-server
. CDK generates a unique 38 character string likefisstackeksclusterchartlitmuse35dacde
which in conjunction with the type extension-server
exceeds the 63 character limit.How to reproduce
Create EKS Cluser and add this