⚠ This page documents a version of Oakestra which is not the latest stable. Please refer to the latest docs for a current version.
Advanced Cluster Setup
Having one root orchestrator and one cluster orchestrator on one device is a great way to start using Oakestra, but the true power of the system lies in its federated architecture.
This guide will walk you through deploying the stand-alone Oakestra components so you can mix and match them to compose the infrastructure you need. Remember, for a valid Oakestra installation, you need at least: 1x Root Orchestrator 🌳, 1x Cluster Orchestrator , and 1x Worker Node 🦾.
Then, you can add additional Cluster Orchestrators and Worker Nodes to extend the infrastructure as much as you need.
For example:
Your Root 🌳 and Cluster Orchestrator can run in a single node, and then you connect 10 more servers, installing a Worker Node 🦾 in each one of them.
Your Root Orchestrator 🌳 runs in a standalone device, and then you deploy three Cluster Orchestrators 🌳 + Worker Node 🦾 on 3 separate servers. This way, you have one root and 3 clusters of 1 node each.
Your Root Orchestrator 🌳 runs in a standalone device; you connect and install a Cluster Orchestrator on 3 separate machines (respectively 𝞨, 𝞫, and 𝞱). Finally, you install the Worker Node 🦾 on 30 other Linux machines, and you connect 20 of them to 𝞨, 5 of them to 𝞫, and an additional 5 to 𝞱.
You probably got the point by now! Let’s learn how you can compose the infrastructure according to your needs.
Here, the instructions on how to install the standalone components and where to install them are up to you.
The Root Orchestrator component will manage your clusters. You will interact with the root orchestrator to deploy and manage your applications via the oak terminal command, the dashboard or the APIs.
Install the oak cli in your root orchestrator machine:
curl -sfL oakestra.io/oak.sh | bash
Install the Root orchestrator using:
oak install root alpha-v0.4.410
What is this doing? This script downloads the required files to the directory ~/.oakestra/root_orchestrator. From there it executes the root orchestrator using docker compose.
You can deploy a cluster orchestrator in the same machine as your root or in a separate machine. You can create as many clusters as you need by deploying multiple cluster orchestrators (each one on a different machine). This component will manage the worker nodes and reports to the root orchestrator aggregated information about the cluster status.
Install the oak cli in your root orchestrator machine:
curl -sfL oakestra.io/oak.sh | bash
Configure your Cluster Orchestrator name and the address of the root machine.
oak config set root_orchestrator_address <IP OF ROOT ORCHESTRATOR> #Only if different than current machineoak config set cluster_name <UNIQUE NAME FOR YOUR CLUSTER>
Install and startup your cluster
oak install cluster alpha-v0.4.410
Be carefull: this install script asks you to confirm the address, name and position of your cluster, the name of the cluster and the geographical coordinates in the format latitude,longitude,radius. The radius is in meters. By default it guesses your location but you can change it during installation. If you agree with the defaults, just press ENTER
What is this doing? This script downloads the required files to the directory ~/.oakestra/cluster_orchestrator. From there it executes the root orchestrator using docker compose.
You can register as many cluster orchestrators with the root orchestrator as you would like. Repeat the above command on a new device and with a unique Cluster Name and Cluster Location.
If you have at least a running Root Orchestrator and at least one Cluster Orchestrator you can add as many new worker nodes to each cluster as you need.
For each worker node you want to install, repeat the following procedure:
Install the oak cli in your worker node machine:
curl -sfL oakestra.io/oak.sh | bash
(Optional) Set the root orchestrator address if that’s running on a different machine
oak config set root_orchestrator_address <IP OF ROOT ORCHESTRATOR> #Only if different than current
Install and run your worker node:
oak install worker alpha-v0.4.410
At the end of the installation, it will ask you the following:
To select the cluster orchestrator where you want to attach this worker node
If you want to start up the worker node right away (y/N)
If you decide to not start your worker node right away, you can run it in background later on running:
oak worker -d
N.b. The cluster selector during the installation shows you the default IP address of the cluster detected by the root orchestrator.
If within your cluster network, you prefer an internal address where the cluster orchestrator is still reachable, check Advanced Installation Options
Network Configuration
If you run into a restricted network (e.g., on a cloud VM) you need to configure the firewall rules and the NetManager component accordingly. Please refer to: Firewall Setup
