Rendered manifests contain a resource that already exists. Could not get information about the resource: resource name may not be empty

I think you have to check if you left off any resource without - name:

I had the same issue. In values.yaml I had

name:

and in deployment.yaml I tried to access this "name" via {{ .Values.name }}. I found out, that {{ .Values.name }} doesn't work for me at all. I had to use {{ Chart.Name }} in deployment.yaml as the built-in object. ref: https://helm.sh/docs/chart_template_guide/builtin_objects/

If you want to access the "name", you can put it into values.yaml for example like this:

something: name:

and them access it from deployment.yaml (for example) like {{ .Values.something.name }}

I suppose there is already the same resource existing in your namespace where you are trying to install or in your helm chart you are trying to create the same resource twice.
Try to create a new namespace and try helm install if you still face the issue then definitely there is some issue with your helm install.

I faced the same error, the fix was to correct the sub-chart name in values.yaml file of main chart.

Best bet would be to run helm template . in the chart directory and verify that name or namespace fields are not empty. This was the case with me atleast.

Most likely, one of the deployments you removed left behind a clusterole.

Check if you have one with kubectl get clusterole

Once you find it, you can delete it with kubectl delete clusterrole <clusterrolename>

Had same error message. I solved the problem using helm lint on the folder of a dependecy charts that I just added. That pointed me to some bad assignment of values. Beware: helm lint on the parent folder didn't highlight any problem in the dependencies folders.