Get started with Realm

Install

Linux / macOS

curl -fsSL https://realm.bitomia.com/install.sh | sudo sh

Windows (PowerShell)

irm https://realm.bitomia.com/install.ps1 | sudo iex

The install location can be overridden with the REALM_INSTALL_DIR environment variable.

Start the agent

On every machine you want to manage, start the agent with:

realm agent start

By default it reads config.yaml from the current working directory. Point it at another file with -c/--config:

realm agent start --config /etc/realm/config.yaml

The agent serves its REST API on 0.0.0.0:9000 unless you override it in the agent section of the config file:

agent:
  listen_address: 0.0.0.0
  listen_port: 9000

That address is what the CLI talks to, so it must match the url you give the node in the nodes section below.

Configure the cluster

A realm config file specifies the nodes of a cluster and the loads that can be deployed on each node.

Lets start with a basic configuration.

nodes:
  lab1:
    url: http://localhost:9000
    driver: linux

loads:
  hello:
    node: lab1
    driver: container
    driver_config:
      image: docker.io/library/hello-world:latest

This config specify that our cluster is composed by just one node lab1 that uses the linux node driver, and this node is running a realm agent on http://localhost:9000 (more on agents later).

It also specify a hello load that uses the container driver with a basic hello-world image settings. This load is meant to be deployed on the lab1 node.

So to summarize the basic primitives:

• Nodes: different cluster nodes where a realm agent is running
• Loads: loads meant to be deployed and control per each node

Now with this configuration you can command your cluster using the realm CLI:

realm n state lab1