This document captures information about using Helm in specific Kubernetes environments.
We are trying to add more details to this document. Please contribute via Pull Requests if you can.
Helm is tested and known to work with minikube. It requires no additional configuration.
Hyperkube configured via scripts/local-cluster.sh is known to work. For raw
Hyperkube you may need to do some manual configuration.
Google's GKE hosted Kubernetes platform enables RBAC by default. You will need to create a service account for tiller, and use the --service-account flag when initializing the helm server.
See Tiller and role-based access control for more information.
Kubernetes bootstrapped with kubeadm is known to work on the following Linux
distributions:
- Ubuntu 16.04
- Fedora release 25
Some versions of Helm (v2.0.0-beta2) require you to export KUBECONFIG=/etc/kubernetes/admin.conf
or create a ~/.kube/config.
Helm requires that kubelet have access to a copy of the socat program to proxy connections to the Tiller API. On Container Linux the Kubelet runs inside of a hyperkube container image that has socat. So, even though Container Linux doesn't ship socat the container filesystem running kubelet does have socat. To learn more read the Kubelet Wrapper docs.
Helm works straightforward on OpenShift Online, OpenShift Dedicated, OpenShift Container Platform (version >= 3.6) or OpenShift Origin (version >= 3.6). To learn more read this blog post.
Helm Client and Helm Server (Tiller) are pre-installed with Platform9 Managed Kubernetes. Platform9 provides access to all official Helm charts through the App Catalog UI and native Kubernetes CLI. Additional repositories can be manually added. Further details are available in this Platform9 App Catalog article.
Helm (both client and server) has been tested and is working on Mesospheres DC/OS 1.11 Kubernetes platform, and requires no additional configuration.