Awesome

kube_dist

Erlang Distribution inside Kubernetes Cluster

With kube_dist properly configured, name of erlang node is set to ${CONTAINER_NAME}@${POD_NAME}.

Moreover, you may use Kubernetes's Service as your process registry.

gen_server:call({via, kube_endpoints, ServiceName}, Request)

Request would be sent to one randomly chosen pod listed in the endpoints list. If ServiceName is an atom, it would be sent to the container with same name. Or ServiceName could a tuple of two atom {ServiceName, ContainerName}, it would be sent to the container which name is specified in the second element of the tuple.

Installation

add kube_dist dependency to your project's rebar.config

{deps, [kube_dist]}.

set start_distribution to false and enable httpc service of inets

[
  {kernel,
   [{start_distribution, false}]
  },
  {inets,
   [{services, [{httpc, []}]}]
  }
].

disable epmd daemon and set epmd_module to kube_epmd and proto_dist to kube_tcp

-start_epmd false
-epmd_module kube_epmd
-proto_dist kube_tcp

grant permission of get to pods/status and watch to endpoints to the service account of the pods

---

apiVersion: v1
kind: ServiceAccount
metadata:
  name: erlang-node

---

apiVersion: rbac.authorization.k8s.io/v1beta1
kind: Role
metadata:
  name: erlang-node
rules:
- apiGroups: [""]
  resources: ["pods/status"]
  verbs: ["get"]
- apiGroups: [""]
  resources: ["endpoints"]
  verbs: ["watch"]

---

apiVersion: rbac.authorization.k8s.io/v1beta1
kind: RoleBinding
metadata:
  name: erlang-node
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: Role
  name: erlang-node
subjects:
- kind: ServiceAccount
  name: erlang-node

and set environment variable CONTAINER_NAME to container's name and add a distribution port named ${CONTAINER_NAME}-dist

spec:
  serviceAccountName: erlang-node
  containers:
  - name: foo
    env:
    - name: CONTAINER_NAME
      value: foo
    ports:
    - containerPort: 4370
      name: foo-dist

License

kube_dist is released under Apache 2 License. Check LICENSE file for more information.