Skip to content

Latest commit

 

History

History
159 lines (114 loc) · 4.91 KB

README.md

File metadata and controls

159 lines (114 loc) · 4.91 KB
Raw

RabbitMQ Kubernetes Operator

Build Status Coverage Status

This project is not mainted anymore. I joined to the Kubernetes RabbitMQ team, so consider to use the official RabbitMQ cluster Operator: https://github.com/rabbitmq/cluster-operator

The easier way to deploy a RabbitMQ Cluster on Kubernetes.

Current Features:

  • Automatic RabbitMQ service creation
  • Configure the right policies to run the Cluster
  • Easy storage-class configuration for persistent the data

Work in progress features:

  • Handle the node remove from the cluster, avoiding to lose the the messages

Parameters

Name Description Values Default
replicas Number of the nodes Integer (1 to x) 1
serviceDefinition Service creating Internal = Auto create the service External = it does not create the service Internal
persistentVolume.storageClass k8s Storage Class definition, if "" won't use the persistentVolume configuration string ""
configMap Config Map name string rabbitmq-config

Install the RabbitMQ Operator

  • Install the operator from Docker-Hub:
kubectl apply -f https://github.com/Gsantomaggio/rabbitmq-operator/releases/download/v0.7-alpha/rabbitmq-operator-lastest.yaml

Deploy RabbitMQ with the Operator

Inside the directory config/samples you can find the deploy examples. The examples are built with kustomize.

Use the command: kubectl apply -k to deploy it.

Localhost developing

For developing purpose you can use config/samples/overlays/developing, so:

kubectl apply -k config/samples/overlays/developing

By setting serviceDefinition = External It creates a custom Service with nodePort configuration, so it can be used in local configuration without load-balancers

Standard deploy

For the standard purpose you can use config/samples/overlays/testing, so:

kubectl apply -k config/samples/overlays/testing

Install the RabbitMQ Operator with Helm3

Add the repo:

 helm repo add alpha https://github.com/Gsantomaggio/rabbitmq-operator/releases/download/v0.7-alpha/

Update the repo list:

helm repo update

Search the package (optional):

helm search repo rabbitmq
NAME                                    CHART VERSION   APP VERSION     DESCRIPTION
alpha/rabbitmq-operator-helm            0.2.7           0.7             Helm for the RabbitMQ Operator

Install the package:

helm install  rabbitmq alpha/rabbitmq-operator-helm

Localhost developing using Kind

Kind is a tool for running local Kubernetes clusters using Docker container "nodes".

Create the Kind cluster:

kind create cluster --config utils/kind/kind-cluster.yaml

The kind-cluster.yaml configuration creates a localhost binding:

 extraPortMappings:
  - containerPort: 31672
    hostPort: 15672
  - containerPort: 30672
    hostPort: 5672

Then you can use the localhost developing, the service.yaml exposes the AMQP and HTTP ports

spec:
  type: NodePort
  ports:
   - name: http
     protocol: TCP
     port: 15672
     targetPort: 15672
     nodePort: 31672
   - name: amqp
     protocol: TCP
     port: 5672
     targetPort: 5672
     nodePort: 30672

So you can easly use it in http://localhost:15672 and amqp://localhost

See the Check the Installation section to test it

Build for source

Requirements:

git clone https://github.com/Gsantomaggio/rabbitmq-operator.git
cd rabbitmq-operator
make

Check the Installation

Describe:

kubectl describe rabbitmq --all-namespaces

Running Pods:

kubectl get pods --all-namespaces | grep rabbitmq
NAMESPACE                  NAME                                                   READY   STATUS    RESTARTS   AGE
ns-developing              rabbitmq-op-developing-0                               1/1     Running   0          22m
ns-developing              rabbitmq-op-developing-1                               1/1     Running   0          20m
rabbitmq-operator-system   rabbitmq-operator-controller-manager-6b695f98d-jfk7j   2/2     Running   0          23m

Project status

The project is still experimental, not ready for production yet.